Rajesh Jain

Published February 3, 2026

1

Why AdWaste Is Structural, Not Accidental

I have discussed NeoMarketing in numerous essays in recent times. Each essay builds on earlier ones and incrementally improves the framework.

Maya’s Dashboard

Maya is the CMO of a mid-sized D2C brand. Her dashboard looks busy — and on the surface, healthy. Campaigns are launching on schedule. The CDP is stitched together. Journeys are flowing. AI is optimising subject lines, send times, and offers in real time.

Yet three numbers refuse to cooperate.

Customer acquisition cost is up 40% over two years. Reacquisition now accounts for 65% of performance spend. Retention is flat, despite more than $2 million invested in martech.

Maya has done everything the playbooks prescribe. She has modern tools, a capable team, and more data than ever before. But customers keep fading, budgets keep rising, and contribution margins keep shrinking.

She’s not alone. Across industries, CMOs face the same pattern. More sophistication, same decay. More spend, same leakage. More tools, same outcome.

This isn’t a failure of effort. Or intelligence. Or tooling.

Maya doesn’t have an execution problem. She has an architecture problem.

**

The AdWaste Loop

The problem shows up as a loop most marketers recognise instinctively, even if they’ve never named it:

Acquire → Ignore → Drift → Reacquire → Repeat.

A customer is acquired at high cost. They engage briefly. Then nothing sustains the relationship. They drift quietly out of view. Months later, the brand pays Google or Meta to “acquire” them again — often without realising they already had them.

This is not growth. It is buying back your own customers at auction.

Across industries, 60–70% of so-called acquisition spend is actually reacquisition. That money doesn’t build new relationships; it compensates for ones that were allowed to decay.

That is AdWaste — and it behaves less like inefficiency and more like a recurring tax. A revenue tax paid to Google, Meta, and marketplaces — funded by marketing’s failure to keep customers in the first place.

And the tax keeps rising. Because the underlying architecture guarantees decay.

**

The Three Structural Failures

AdWaste persists not because marketers are careless, but because marketing systems were built on assumptions that no longer hold. Three structural failures sit beneath the loop.

Failure	What Broke	Symptom
Intelligence Gap	Humans can’t do N=1	Segments decay, Best customers fade undetected
Incentive Gap	Vendors paid for activity	No accountability for retention outcomes
Attention Gap	Push without memory	Inbox forgets, reacquisition becomes inevitable

The Intelligence Gap. Humans cannot manage continuous N=1 relationships at scale. Segments decay. Campaigns lag reality. Best customers fade because no system is watching them closely enough, often enough.

The Incentive Gap. Most vendors are paid for activity, not outcomes. Emails sent, messages delivered, journeys launched — revenue accrues regardless of whether customers stay or leave. Retention is optional; volume is not.

The Attention Gap. Marketing remains push-driven in a world that rewards pull. Channels forget. Engagement doesn’t compound. Customers fade silently until they reappear — expensively — in an ad auction.

These failures reinforce one another. Together, they make AdWaste inevitable, regardless of how advanced the tooling becomes.

**

Why Martech Can’t Fix This

The instinct is to solve these problems with better martech. Better AI. Better personalisation. Better targeting.

But martech optimises inside the broken structure. It does not replace it.

Backend memory is not experiential memory. Campaigns are not continuity. Channels are not relationships.

Martech didn’t fail at retention. It was never designed for it.

Fixing AdWaste does not require better marketing. It requires a new operating system.

2

Three A’s as the New Operating System

If AdWaste is structural, then the solution must be architectural.

NeoMarketing is not an upgrade to martech. It is a replacement for the assumptions martech was built on — a system designed so AdWaste becomes structurally impossible.

**

The Framework

AGENTIC → Never Lose Customers
ALPHA → Never Pay Fixed
ATTENTION → Never Pay Twice

Each pillar eliminates one structural failure. Together, they replace acquisition-first marketing with continuity-first growth.

**

AGENTIC — Never Lose Customers

Structural failure solved: the Intelligence Gap.

The maths of modern marketing is unforgiving. Ten thousand customers, multiple channels, daily decisions, real-time context — no human team can continuously optimise at N=1. Segmentation is a compromise, not a solution.

NeoMarketing replaces campaign-centric execution with Agentic systems.

M-Agents are autonomous AI agents for marketing ops that monitor intent, detect early disengagement, and orchestrate personalised actions continuously — not in batches, not in campaigns, but as a living system.

BrandTwins, created via the TwinFactory, act as persistent customer-side advocates. They learn preferences, protect attention, and ensure relevance before customers drift.

The result is not better campaigns. It is the end of campaigns as the primary unit of marketing.

Best customers stay Best. Fade is detected before it becomes churn.

Agentic doesn’t automate campaigns. It eliminates the need for them.

**

ALPHA — Never Pay Fixed

Structural failure solved: the Incentive Gap.

Traditional martech pricing is input-based. Send more, pay more. Use more features, pay more. Success or failure is irrelevant to the vendor’s revenue.

NeoMarketing introduces Alpha economics.

Pricing shifts to an outcome-based model:

• Beta: a modest fixed baseline
• Alpha: performance-linked upside
• Carry: long-term profit participation

This is paired with Progency — an operating model that combines product, agents, and agency-like strategic services (via Martech Growth Engineers), paid for results rather than effort.

When vendors are paid for retention and profit, behaviour changes. Accountability becomes unavoidable.

Alpha doesn’t just change pricing. It changes who is responsible for results.

**

ATTENTION — Never Pay Twice

Structural failure solved: the Attention Gap.

Email and owned channels lost their magnet. Each interaction resets. Engagement never compounds. Customers fade quietly — and then get reacquired at full price.

NeoMarketing fixes this at the channel level.

At its core is the Attention Processing Unit (APU) — the primitive that gives the inbox memory.

• Mu in the subject line signals accumulated value
• Magnets create repeatable engagement
• ActionAds monetise attention without disruption
• Mu Ledger makes progress visible and redeemable

APU is deployed in two ways:

• NeoBoost, embedded into existing emails on any ESP, prevents fade among Best customers
• NeoMails, APU-native daily emails, recover Rest/Test customers

NeoNet is the ad network that enables cooperative, deterministic recovery without auction taxes.

The outcome is owned attention that compounds. Reacquisition becomes unnecessary.

Attention doesn’t fix individual emails. It fixes the channel.

**

The Flywheel

These pillars are not independent.

Agentic prevents loss. Attention compounds engagement. Alpha enforces discipline.

Remove any one, and AdWaste returns.

Together, they form a closed system where customers are acquired once, retained continuously, and monetised without leakage.

This is not better martech. This is what replaces it.

3

Platform, Not Products

NeoMarketing is not a product bundle. It is a platform built on primitives — and primitives compound.

**

Products, Primitives, Infrastructure

Layer	Examples	Role
Products	NeoBoost, NeoMails, Progency	Entry points
Primitives	APU, BrandTwins, Alpha	Structural logic
Infrastructure	M-Agents, NeoNet, Mu Ledger	Invisible power

Products can be copied. Primitives cannot be shortcut. Infrastructure creates moats.

This is why adding “AI personalisation” or “better journeys” does not replicate NeoMarketing. The value is not in features; it is in how the system is wired.

The question is not “which product should I buy?” It is “which layer do I need to build on?”

**

New Metrics for a New System

A new architecture demands new measures.

Open rates measure moments. CAC measures spending. Campaign ROAS measures tactics.

NeoMarketing tracks continuity and economics:

Old Metric	Measures	NeoMarketing Metric	Measures
Open Rate	Moments	Attention Retention Rate	Engagement that compounds
CAC	Spending	Reacquisition Ratio	Waste made visible
Campaign ROAS	Tactics	N=1 Live Ledger	Relationship profitability

ARR doesn’t create memory. Memory creates ARR.

**

What Zero AdWaste Actually Means

Zero AdWaste is not lower CAC or better ROAS.

It is:

• Customers acquired once, retained continuously
• Attention earned, not rented
• Owned channels that compound
• Marketing as a profit engine, not a cost centre

Zero AdWaste is not a metric. It is an architectural state.

**

Maya’s New Dashboard

Six months later, Maya’s dashboard tells a different story.

Reacquisition spend is down 35%. ARR is measured — and rising. Customer P&L is visible for the first time. Marketing contribution margin is positive.

She didn’t work harder. She didn’t optimise faster.

She changed the architecture.

Maya stopped fighting AdWaste. She made it structurally impossible.

**

Summary

The next decade of marketing will not be about better ads.

It will be about not needing them.

Traditional Martech: Lose customers. Pay twice. Repeat.

NeoMarketing: Never Lose Customers. Never Pay Fixed. Never Pay Twice.

**

NeoMarketing — The Zero AdWaste Platform

AGENTIC → Never Lose Customers
ALPHA → Never Pay Fixed
ATTENTION → Never Pay Twice

Pay Once. Profit Forever.

4

A Transformation and A Startup

Who builds NeoMarketing? If it is the answer to AdWaste, and if it cannot be replicated by adding features, then how does the industry actually get there?

The answer is uncomfortable but unavoidable: NeoMarketing requires two parallel efforts, not one. A transformation and a startup. Moving at different speeds, with different economics, toward the same destination.

**

Why Incremental Change Fails

When NeoMarketing is first described, a predictable objection arises: Can’t existing martech platforms just add these capabilities?

The answer is no — not because martech companies lack talent or technology, but because NeoMarketing breaks too many foundational assumptions to be layered on top.

You cannot simply “add” Agentic systems to an organisation designed around campaigns and quarterly planning. You cannot “pilot” outcome-based pricing inside a revenue model built on fixed SaaS fees and volume incentives. And you cannot “experiment” with inbox-level attention inside roadmap cycles optimised for enterprise feature delivery.

The Three A’s — Agentic, Alpha, Attention — are not features. They are architectural shifts. Each one changes who does the work, who bears the risk, and who is accountable for outcomes.

NeoMarketing cannot be built by doing martech better. It requires doing something different.

**

Martech 2.0 — The Transformation Engine

This does not mean traditional martech companies are obsolete. In fact, they are uniquely positioned to build part of the NeoMarketing OS.

They already have critical assets: customer relationships, data infrastructure, AI foundations, delivery expertise, and trust earned over years. But these assets must be rewired around retention and outcomes — not activity.

This is the transition from Martech 1.0 to Martech 2.0.

Martech 1.0	Martech 2.0
Campaign-centric	Customer-centric
N=Many Segments	N=Few and N=1 via Agents
Fixed SaaS fees	Alpha-based economics
Tools vendor	Growth partner

The shift is not cosmetic.

In Martech 2.0, AI is no longer used to optimise campaigns, but to power M-Agents and BrandTwins that continuously sense intent, detect early fade, and act on behalf of both brand and customer.

Just as importantly, Martech 2.0 abandons the safety of fixed pricing. Alpha economics demand a cultural shift — from selling software to sharing accountability. Revenue becomes tied to retention, growth, and profit — not emails sent or journeys launched.

This is where the Progency model emerges: product, agents, and strategic services combined — paid for results, not effort.

This is hard. It requires cultural change, not just technical change. But it is possible — for companies willing to transform rather than merely upgrade.

Martech 2.0 doesn’t sell tools. It shares responsibility.

**

Neo — Why Attention Must Be Built as a Startup

If Agentic and Alpha can be built through transformation, Attention cannot.

Attention is fundamentally different. It cannot be “evolved into.” It must be born new.

The Attention stack operates under a different physics:

• It depends on network effects — Mu, ActionAds, and cooperative inventory grow in value only as participation increases.
• It involves three actors simultaneously: consumers, brands, and advertisers.
• It runs on habit formation, not procurement cycles.
• It must be ESP-agnostic, working across existing platforms rather than reinforcing one. That positioning is impossible for a traditional martech company protecting its core business.
• And it demands speed, experimentation, and tolerance for failure that transformation-led organisations struggle to sustain.

The Attention Processing Unit (APU) is a primitive that does not exist in today’s martech stack. NeoBoost and NeoMails are carriers for that primitive. NeoNet is not an adtech clone, but a cooperative alternative to auctions — deterministic, identity-based, and aligned with retention.

This is not a roadmap extension. It is a platform creation problem.

This is why Attention must be built as a separate entity. New team. New economics. New ambition. Startup energy applied to a problem that incumbents cannot solve from within.

Attention cannot be retrofitted into martech. It must be built as a platform — with startup energy and network economics.

**

One Vision, Two Engines

Put together, the picture becomes clear.

Engine	Builds	Solves
Martech 2.0	Agentic + Alpha	Intelligence + Incentives
Neo	Attention	Memory + Magnetism

Neither engine works alone.

Agentic without Attention still leaks customers into stateless channels. Attention without Agentic lacks the intelligence to prevent fade. Alpha without both collapses back into fixed-fee SaaS.

Together, they form a closed system. One transforms what must be trusted. The other invents what must scale. They move at different speeds, but they are architecturally integrated.

NeoMarketing is not one company’s product. It is an operating system built by two engines moving at different speeds.

**

This is not one vendor’s roadmap. It is a category reset.

The martech industry does not need more tools, more dashboards, or more AI features layered onto old assumptions. It needs a new architecture — and the courage to build it in two ways at once.

Transform what must be trusted. Start fresh where new models and networks are required.

That is how NeoMarketing moves from framework to infrastructure.

5

Graphical View

Here is a ChatGPT graphic which captures the NeoMarketing vision.

And this is Claude’s take.

 

Published February 2, 2026

1

Two Inboxes

Ria opens her inbox the way most people do now — not with curiosity, but with mild dread.

It’s 8:47am. She’s on the metro, coffee in one hand, phone in the other. There are 47 unread emails. Most arrived overnight. Subject lines stack on top of each other, each shouting in its own way. Flash Sale. Last Chance. Only Today. Still Thinking? We Miss You!

Each email feels like a stranger tapping her shoulder, asking for attention without acknowledging the last interruption.

She scrolls without tapping. The senders blur together. Fashion. Food. Travel. Finance. Brands she once bought from, browsed from, signed up for — now competing for the same exhausted glance. Nothing connects. No sense of history. No reason to linger. Just urgency piled on urgency.

She selects all, deletes all.

The inbox is cleared — not because anything meaningful happened, but because it needed to be emptied. That’s what the inbox has become: something to manage, not something to visit. A graveyard of unread intent. A place you clean up so you can move on.

Brands remember Ria. Her inbox doesn’t.

**

Three months later, same metro, same coffee. But something has changed.

Ria opens her inbox again — and this time, she’s looking for something.

The unread count is higher than before, but she doesn’t rush to clear it. Her eye goes to a single subject line near the top:

µ.1247 | Did you get yesterday’s prediction right?

She taps.

Inside, the email doesn’t demand anything. It continues something.

Yesterday’s prediction is revealed — she had guessed the Nifty would close up. It did. She was right. A small reward appears: +15 Mu. Her Mu balance updates: 1262. A streak badge shows Day 14. She’s opened an email from this brand — or one like it — for two weeks straight. Not because anyone asked her to. Because something was waiting.

She scrolls. A quick poll: “Weekend vibe — brunch or hiking?” She taps brunch. Another +5 Mu. A product carousel appears — not random, but relevant. She’d browsed hiking boots last week; now she sees them with a note: “Still thinking about these?” She swipes past. Not today.

At the bottom, an ActionAd from a skincare brand she’s never heard of. But it’s not irrelevant noise — it’s a product for the dry skin she mentioned in a quiz two weeks ago. She taps. Samples added to cart.

Fifty-eight seconds. Done.

She closes the email. The inbox doesn’t feel empty. It feels alive. Not busy — alive. There’s a sense that something is in motion, that today’s interaction will matter tomorrow.

She doesn’t need to remember what she did yesterday. The inbox does.

Later that evening, another email arrives. Different brand. Different category. But the experience feels familiar. The same quiet continuity. The same sense that this isn’t a one-off interruption — it’s part of an ongoing thread.

Ria doesn’t think about loyalty. She doesn’t think about engagement. She doesn’t think about brands at all.

She thinks: I’ll check again tomorrow.

**

That’s the difference.

Nothing dramatic changed. The brands didn’t suddenly get smarter. The offers weren’t louder. The copy wasn’t more clever. No one begged harder for attention.

What changed was the inbox itself.

It stopped being a pile of disconnected messages and became a place where things carried forward. Where time mattered. Where attention accumulated instead of evaporating.

The inbox stopped being something to clear.

It became something to return to.

**

This essay is about what made that shift possible — not a new campaign strategy, not better copywriting, not smarter segmentation.

Something more fundamental.

A layer that didn’t exist before. A primitive that connects emails to each other, independent of the brands that send them. A reason to return that lives in the inbox itself, not in any single message.

The inbox lost its magnet years ago, when personal communication migrated to WhatsApp, iMessage, and Slack. What remained was brand noise — each sender shouting alone, no one listening.

What Ria experienced wasn’t a better email.

It was email with its magnet restored.

Related essays: The Magnetic Inbox and 4 stories.

2

The Channel That Lost Its Magnet

When people talk about email’s decline, they usually blame brands.

Too many emails. Too much frequency. Too little relevance. Bad copy. Worse timing. Lazy segmentation. The conclusion is always the same: brands abused the channel, and customers tuned out.

That story is comforting — and incomplete.

Email didn’t weaken because brands suddenly became incompetent. It weakened because the channel itself lost the one thing that made it magnetic in the first place.

**

For most of its early life, email wasn’t primarily a marketing channel. It was a personal one.

The inbox mattered because it carried people. Messages from friends. Replies from colleagues. Notes that required response. The first thing you did each morning was check email — not out of obligation, but anticipation. Something might be waiting. Someone might have written.

That was the magnet: personal communication. The human pull that made the inbox worth returning to. Brands simply piggybacked on that gravity.

Then something structural changed.

Personal communication migrated.

One by one, the most valuable inbox messages left email and moved elsewhere — to WhatsApp, iMessage, Slack, Telegram. Conversations became faster, richer, more immediate. Group chats replaced threads. Voice notes replaced long replies. The urgent and the intimate moved elsewhere.

And when personal communication left, email’s magnet left with it.

Email didn’t die. It hollowed out.

What remained was not conversation, but content. Brand emails. Promotional messages. Transactional updates. Newsletters nobody asked for. Receipts. Password resets. Each sender speaking independently, each demanding attention, none providing a reason to return tomorrow.

**

This distinction matters.

A channel with a magnet pulls people back even when nothing urgent is happening. A channel without one must constantly push, shout, and escalate just to be noticed.

Once email lost its natural pull, decay became the default state.

Brands didn’t cause this. They inherited it.

Faced with a channel that no longer drew people in on its own, brands did the only thing they could: send more. More campaigns. More reminders. More “personalisation.” More urgency layered on top of urgency.

But volume can’t replace gravity.

The industry poured millions into making brand emails better — better subject lines, better timing, better targeting, better AI. None of it addressed the structural problem.

Email’s magnet was never content. It was the reason to return. And that reason had left the building.

**

Consider the contrast with WhatsApp.

Why do people check WhatsApp dozens of times a day? Not because brands send great messages. Because friends are there. Group chats are active. Something is always waiting — a reply, a reaction, a conversation in progress.

WhatsApp has a magnet: P2P communication and group dynamics. The magnet is not content quality; it’s anticipation. The channel pulls people back independent of any commercial content.

Email has no equivalent. Not anymore.

Each brand email arrives as an island. No connection to the previous message. No anticipation of the next. No shared sense of progress or accumulation. Even the best-crafted email must fight from zero every time.

Hope is not a magnet.

**

This explains a pattern that has puzzled marketers for years.

Why does engagement decay even when emails are “personalised”? Why do Best customers fade despite sophisticated automation? Why do open rates decline even as targeting improves?

Because the problem isn’t the emails. The problem is the channel.

Every email begins the relationship again, as if nothing came before. Opens reset. Clicks evaporate. Attention leaks. Engagement never compounds because the channel stopped remembering.

This is the mistake most analyses make: they treat email’s decline as a brand problem to be solved with better tactics. But you cannot fix a channel-level failure with sender-level optimisation.

You cannot optimise your way to a magnet.

No amount of smarter copywriting, finer segmentation, or AI-driven personalisation can restore a magnet that no longer exists. Those tools can improve individual messages. They cannot make the inbox itself worth returning to.

**

Here’s the hard truth:

Email, as a channel, is structurally disadvantaged — not because it’s old, but because its original source of gravity migrated elsewhere. WhatsApp kept its magnet. Email lost its.

And yet email remains the most valuable owned channel brands have. It’s universal. It’s permissioned. It’s persistent. It reaches customers directly, without algorithmic interference, without platform taxes. The infrastructure is sound. The channel is intact.

What’s missing is the magnetism.

Email didn’t fail because brands abused it. It weakened because its original magnet — personal communication — left the channel.

Restoring email doesn’t start with better emails. It starts with restoring the magnet.

And that requires something brands cannot build on their own — a layer that reconnects messages, creates continuity, and gives people a reason to return to the inbox itself, not just to any single sender.

The next section explores why brand-level improvements can’t solve a channel-level problem — and what kind of solution actually can.

3

Why Brand Emails Can’t Fix a Channel Problem

Once you accept that email lost its magnet at the channel level, a second truth becomes unavoidable: no individual brand can fix this on its own.

And yet, that’s exactly where the industry has spent the last decade trying.

Brands have invested heavily in making their emails better. Customer Data Platforms stitch together behaviour across touchpoints. Journey builders orchestrate increasingly complex flows. AI engines optimise subject lines, timing, and content in real time. Personalisation is deeper than ever. Automation is more sophisticated than ever.

This is real progress — but it’s progress of a specific kind.

Call it backend memory.

Backend memory is the system’s ability to remember about the customer: what they browsed, what they bought, what they opened, where they dropped off. It lives in databases. It benefits brands. And it’s excellent at improving targeting and efficiency.

What backend memory cannot do is make the inbox itself worth returning to.

**

From the brand’s point of view, things look coherent. The customer has a profile. A history. A place in a journey.

From the customer’s point of view, none of this is visible. Each email still arrives in isolation, disconnected from the last and unrelated to the next.

The brand remembers everything. The inbox reflects nothing.

This is the island problem.

Brand A sends an email. Brand B sends an email. Brand C sends an email. Each may be personalised, well-timed, and relevant in isolation. But the customer doesn’t experience isolation — they experience the inbox as a whole.

And that whole is fragmented.

There is no continuity across senders. No accumulation of value. No shared sense that opening one email makes the next more meaningful. Each brand is optimising its own message, but no one is responsible for the experience of the channel.

It’s like every shop on a dying high street improving its own window display. The individual windows get better. The street stays empty. Because the problem isn’t the shops — it’s that no one has a reason to walk down the street anymore.

The inbox is the street. And the street has lost its pull.

**

This is why even “good” emails fail to compound.

A brand can improve its open rate this week. It cannot, on its own, create anticipation for the inbox tomorrow. Because anticipation is not brand-specific — it is channel-specific.

The industry assumed that if every sender got better, the channel would recover. But that logic only works when a channel already has gravity. When the magnet is gone, optimisation turns into escalation. More urgency. More frequency. More noise.

Better subject lines? They help one email compete against other emails — but don’t change whether the inbox gets opened in the first place.

Smarter personalisation? It makes individual messages more relevant — but doesn’t create a reason to return tomorrow.

AI-powered send time optimisation? It finds the best moment to interrupt — but interruption is not attraction.

These tactics optimise the push. They don’t restore the pull.

**

What’s missing is not intelligence. It’s continuity.

Specifically, what’s missing is inbox memory — memory that lives in the experience the customer sees, not just in the systems brands operate. Memory that is visible at the moment of action. Memory that turns isolated interactions into an ongoing thread.

This distinction matters because behaviour only changes when memory is visible.

Think about why other digital experiences create habits. Social feeds show you what you liked before — memory made visible. Games display your streak, your level, your progress — memory made visible. Wallets show your balance, your rewards, your status — memory made visible.

Databases can remember forever, but invisible memory does not form habits. A customer cannot feel a journey they cannot see. They cannot value progress that never appears. They cannot anticipate what the inbox does not signal.

Backend memory helps brands target better. Inbox memory makes the channel worth returning to.

Until now, email has had the first and never the second.

**

That’s why engagement resets with every send. That’s why Best customers fade without triggering alarms. That’s why attention decays even as personalisation improves.

The problem isn’t that brands aren’t remembering. It’s that the inbox isn’t.

This is not a failure of execution. It’s a missing layer.

To restore email’s magnet, you don’t start by asking, “How can brands send better emails?” You ask a different question:

What would make the inbox itself worth coming back to — regardless of who sent the last message?

Answering that requires something email has never had before. Not a campaign. Not a tactic. Not a feature bolted onto individual messages.

It requires a new primitive — one that operates at the channel level, reconnects emails to each other, and makes memory visible where it actually changes behaviour: inside the inbox.

That primitive is what the next section introduces.

4

The Primitive That Restores the Magnet

To close this gap requires more than better emails. It requires a new layer — one that operates at the inbox level, not the message level.

This is where the Attention Processing Unit (APU) comes in.

APU is not a feature. It’s not a widget. It’s not a campaign tactic dressed up in new language.

It’s a primitive — a foundational layer that makes inbox memory possible.

Think of it this way: APU is to attention what a database is to applications. A database doesn’t create the application, but without it, nothing persists. Nothing accumulates. Nothing continues. The database is the layer that makes memory possible.

APU does the same for the inbox.

**

Here’s what APU actually does: it carries visible memory forward from one email interaction to the next — independent of the brand that sends the message.

That last part is crucial.

Until now, every attempt to improve email has been sender-specific. Brand A improves Brand A’s emails. Brand B improves Brand B’s emails. Each operates in isolation. The inbox remains fragmented.

APU works differently. It creates a continuity layer that sits across senders. A customer’s engagement with a coffee brand’s email connects to their engagement with a fashion brand’s email. Progress accumulates. Memory persists. The inbox itself becomes coherent.

This is not coordination between brands. It’s infrastructure beneath them.

**

How does this work in practice?

APU has four components, each serving a specific role in creating inbox-level magnetism:

Mu in the Subject Line — the signal that something is waiting.

When Ria sees “µ.1247” in a subject line, she knows this isn’t just another promotional email. It’s a cue. Something has accumulated. Something can be continued. The subject line becomes a beacon — a reason to open that exists before she even sees the content.

This is fundamentally different from “Don’t miss out!” or “Sale ends tonight!” Those demand attention. Mu signals that attention has been remembered.

Magnets — the reason to engage.

Inside the email, a Magnet provides interaction that pulls rather than pushes. A prediction to resolve. A quiz to complete. A game to continue. Yesterday’s interaction sets up today’s. Today’s sets up tomorrow’s.

Magnets are not brand content. They’re engagement mechanics that work across senders. The coffee brand’s email might have a prediction. The fashion brand’s email might have a quiz. But the Mu earned in one carries to the other. The streak maintained in one counts toward the other.

This is how disconnected emails become connected experiences.

ActionAds — relevance without repetition.

Traditional ads in email are interruptive and often irrelevant. ActionAds are different. They’re targeted based on identity — not cookies, not probabilistic matching, but authenticated, first-party knowledge of who the customer is and what they’ve engaged with.

An ActionAd in a coffee brand’s email might show a skincare product — because the customer mentioned dry skin in a quiz two weeks ago. The targeting is precise. The experience is relevant. And because the system remembers previous interactions, it doesn’t repeat offers already seen or declined.

This has never been possible in email (or any other platform) before.

Mu Ledger — the accumulation made visible.

Every interaction earns Mu. Every Mu is recorded. The balance is visible — in subject lines, inside emails, across the inbox. Progress doesn’t vanish after a click. It accrues.

The Ledger also shows redemption options. Mu isn’t abstract points — it’s value that can be exchanged. This closes the loop: attention is earned, remembered, and rewarded.

**

Now step back and see how these components work together.

The subject line signals continuity. The Magnet earns engagement. ActionAds monetise attention without disrupting it. The Ledger accumulates value. Each component reinforces the others.

But the real breakthrough isn’t any single component. It’s what they create in aggregate:

A reason to return to the inbox itself.

Not to Brand A’s email. Not to Brand B’s email.

Crucially, the continuity does not belong to any one brand — it belongs to the inbox.

The customer opens her inbox knowing that something is waiting — Mu to collect, a streak to maintain, a prediction to resolve. The specific sender matters less than the fact that the inbox now has gravity. Every brand benefits because the channel has regained its pull.

This is the shift from message-level optimisation to channel-level magnetism.

**

APU doesn’t replace what brands already do. It doesn’t compete with CDPs or journey builders or personalisation engines. Those tools still matter for backend memory — for helping brands decide what to send.

APU adds what they cannot provide: memory the customer can see and feel. Continuity that spans senders. A magnet that lives in the inbox, not in any individual message.

The relationship stops being merely stored in a database.

It becomes experienced in the inbox.

**

But a primitive needs carriers. APU doesn’t deliver itself. It needs to be embedded into the emails customers actually receive.

That’s where NeoBoost and NeoMails come in — two deployment modes for the same underlying primitive, each solving a different problem in the attention lifecycle.

The next section explains how they work.

5

One Primitive, Two Carriers

If APU is the primitive that restores magnetism to the inbox, then NeoBoost and NeoMails are simply how that primitive enters the real world.

They are not competing products. They are not alternative strategies. They are deployment modes — two ways to carry the same continuity layer into customer inboxes, depending on where the breakdown in attention has already occurred.

The distinction is straightforward.

	NeoBoost	NeoMails
Target	Best (fading)	Rest/Test (disengaged)
Problem	Retention	Recovery
Email type	Existing emails	New daily stream
ESP	Any ESP	Netcore required

NeoBoost is designed for customers who are still there — opening, clicking occasionally, but slowly fading. These customers don’t need a new stream. They need the emails they already receive to stop resetting the relationship every time. NeoBoost embeds APU into transactional, promotional, and newsletter emails sent through any ESP — no migration required, no platform fees. It prevents decay before it becomes loss.

NeoMails is designed for customers who have already drifted — the large, silent majority who haven’t churned but no longer engage. For them, brands need a fresh surface, a new reason to re-enter the inbox. NeoMails creates an APU-native email stream built entirely around continuity, rewards, and pull. It rebuilds broken relationships.

One preserves attention. The other recovers it.

Together, they close the attention loop.

**

But neither NeoBoost nor NeoMails is the real point.

They are carriers. What they carry is APU — the layer that makes continuity possible across emails, across brands, and across time.

This distinction matters because it reframes how success should be measured.

Traditional email metrics — opens, clicks, CTR — measure moments. They tell you whether a message worked once. They say nothing about whether engagement is compounding or leaking.

What matters in a world with inbox memory is Attention Retention Rate (ARR): the percentage of customers who interact with APU quarter over quarter.

ARR is not a KPI to optimise. It’s a vital sign.

ARR doesn’t create memory. Memory creates ARR.

High ARR means the inbox has regained pull. Customers are returning not because they’re being chased, but because something persists. Low ARR means silent fade is underway, even if individual campaigns look healthy.

Seen this way, ARR becomes a leading indicator of reacquisition spend. When ARR holds, reacquisition drops. When ARR decays, ad budgets quietly rise.

**

This brings us back to the bigger picture.

For twenty-five years, email has been experientially stateless. Each message a fresh start. Each brand an island. Each interaction forgotten the moment it ends. When personal communication left the channel, nothing replaced it.

APU changes that at the channel level.

Mu makes attention worth accumulating. Magnets make return worth anticipating. ActionAds make monetisation additive, not extractive. The Ledger makes progress visible.

The inbox stops being a pile of disconnected messages and becomes a place where things carry forward.

An inbox that pulls, not just pushes. A channel restored. A magnet returned.

**

This is the idea that ties everything together:

The Magnetic Inbox with Memory.

Magnetic — because it earns attention through pull, not push. Customers return because something is waiting, not because someone demanded it.

Memory — because it makes attention compound, not decay. Each interaction builds on the last. Progress persists. Relationships deepen instead of resetting.

**

Remember Ria?

Her inbox used to be a graveyard. Forty-seven unread emails, each shouting independently, none connecting to the others. A place to clear, not check.

Now she opens her inbox looking for something. A prediction to resolve. A streak to maintain. A balance to grow. The brands are still there — but the experience is transformed.

What changed wasn’t the brands. It was the inbox itself.

**

Marketing automation gave brands memory about customers.

APU gives the inbox memory that customers can feel.

Brands remember customers. Now the inbox does too.

Published February 1, 2026

1

Preamble, Three Pillars

The Question We Left Unanswered

The previous essay diagnosed how martech created the $500 billion AdWaste crisis.

Eight structural failures — from input-based pricing to the reacquisition mirage — formed a self-reinforcing system that turned the industry meant to solve retention into adtech’s most valuable upstream partner.

The result:

• 80% quarterly attention churn
• 90% of marketing budgets flowing to adtech
• 70% of that spend going to reacquisition
• Half a trillion dollars in annual waste

The diagnosis ended with a question:

What must fundamentally change?

This essay provides the answer.

**

Why Fixes Won’t Work

The Instinct to Patch

After diagnosis, the instinct is to reach for familiar remedies:

• Better segmentation
• Smarter automation
• More AI features
• Improved dashboards

These feel like progress. They are not.

The Logic of Failure

The eight failures were not isolated mistakes. They were symptoms of three deeper misalignments:

• Misaligned Economics Vendors paid for activity, not outcomes
• Misaligned Capability Tools exceeded human capacity to operate
• Misaligned Attention Push systems in a pull world

Why Optimisation Fails

• You cannot fix input pricing with better dashboards The incentives stay broken
• You cannot fix attention decay with better copy The physics of push guarantees decay
• You cannot fix reacquisition with smarter retargeting Retargeting is reacquisition — the problem itself

Any solution that keeps the same incentives, metrics, and architecture will reproduce the same outcomes.

AdWaste cannot be optimised away. It must be structurally eliminated.

**

Introducing NeoMarketing – the Anti-Martech.

What NeoMarketing Is Not

• “AI marketing” (AI is a means, not an end)
• “Better martech” (it replaces, not improves)
• “Next-gen CRM” (CRM is a subset, not the system)

What NeoMarketing Is

• A replacement operating system
• Built to eliminate AdWaste
• Designed around retention, not reacquisition

The Core Doctrine: Never Lose Customers. Never Pay Twice.

The extension: Pay Once. Profit Forever.

**

The Three Pillars

NeoMarketing rests on three co-equal foundations — the Three A’s:

Pillar	Function	What It Fixes
ALPHA	Economics	Misaligned incentives
AGENTIC	Intelligence	Operational complexity
ATTENTION	Engagement	Push-based decay

Remove any one, and AdWaste returns.

2

The Pillars

ALPHA — The Economic Reversal

What It Fixes: Root Cause #1 (Input Pricing) and #7 (Adtech’s Advantage)

The Diagnosis Recap

Martech failed because it sold activity, not results.

Vendors charged for:

• Messages sent
• Records stored
• MTUs / MAUs / API calls

They got paid whether:

• Customers engaged or ignored
• Retention improved or collapsed
• Attention decayed or compounded

Failure was profitable. Success was optional.

The NeoMarketing Shift

From Input-Based Pricing → Outcome-Based Economics

Alpha realigns incentives by redesigning the commercial model entirely.

The Alpha Structure

Beta (Baseline) A lower fixed component covering operational costs. Ensures function, but not where profits are made.

Alpha (Uplift) Variable compensation tied directly to retention lift, revenue growth, customer recovery, and attention metrics. If customers stay, vendors earn. If customers lapse, vendors lose.

Carry (Shared Upside) Long-term participation in sustained value created. Ensures vendors think in years, not quarters.

What Changes

• Vendors have skin in the game
• Sending messages that don’t work hurts vendor economics
• Retention becomes the primary objective, not an afterthought

Martech Growth Engineers (MGEs)

Alpha economics requires a different kind of engagement.

MGEs are hybrid professionals — part strategist, part operator, part analyst — who work directly with brands to ensure NeoMarketing delivers results.

They are:

• Accountable for growth outcomes
• Compensated on the same metrics as the platform
• The human bridge until AI agents mature

Over time, MGE playbooks become agent playbooks. The knowledge transfers; the alignment persists.

Martech vendors collected rent. NeoMarketing partners earn yield.

**

AGENTIC — The Intelligence Reversal

What It Fixes: Root Cause #2 (Complexity), #3 (Personalisation Mirage), and #4 (Attention Blind Spot)

The Diagnosis Recap

Martech promised “Segment of One.” What it delivered was “Batch and Blast.”

Why?

• Platforms bloated with features: 60-70% went unused
• Teams stayed small: one person managing 500,000 subscribers
• Content creation didn’t scale
• Data files stayed thin
• Attention decay went unmeasured

The result: sophisticated tools producing primitive outputs.

The NeoMarketing Shift

From Tools for Humans → Agents That Execute

The solution is not simpler tools. Simpler tools cannot deliver N=1 at scale. The solution is not more training. Training cannot close a 1:500,000 gap.

The solution is agents — AI systems that operate the complexity humans cannot.

Marketing Agents: The Operational Layer

Agent	Function
Insights Agents	Analyse behaviour, surface anomalies, generate intelligence
Segment/Audience Agents	Create dynamic audiences based on trajectories, not static attributes
Content Agents	Generate and adapt messaging at scale
Shopping Agents	Optimise commerce experience based on individual context

Together, they replace the “armies of specialists” that brands couldn’t afford.

Customer Agents: The Relationship Layer

BrandTwins are AI advocates for individual customers — not segments, not personas.

Each BrandTwin:

• Knows purchase history, engagement patterns, preferences
• Predicts needs and identifies optimal moments
• Detects drift before it becomes dormancy
• Advocates for the customer within the brand’s systems

The architecture:

ArtificialPeople: Foundational consumer world models trained on behavioural patterns across populations.

Twin Factory: The customisation engine that creates and maintains millions of individual BrandTwins by overlaying brand-specific data onto ArtificialPeople foundations.

Live Ledger: Making Attention Visible

Each BrandTwin has a corresponding ledger entry tracking:

• Attention state (engaged, drifting, dormant)
• Trajectory (improving, stable, declining)
• Lifetime value (historical and predicted)
• Intervention opportunities
• Risk signals

The Live Ledger transforms customers from campaign targets to continuously managed assets.

Martech sold Ferraris without drivers. Agentic provides the autonomous driver.

**

ATTENTION — The Engagement Reversal

What It Fixes: Root Cause #5 (Inbox Failure), #6 (Best Customer Bias), and #8 (Reacquisition Mirage)

The Diagnosis Recap

Email should have been martech’s fortress. It became its liability.

• Batch-and-blast trained customers to ignore the inbox
• 80% quarterly attention churn became normal
• The “missing middle” (Rest and Test — 80% of customers) was abandoned
• When they lapsed, the only path back was adtech auctions

Brands paid Google and Meta to reach customers they already owned, because martech had trained those customers to stop listening.

The NeoMarketing Shift

From Push Campaigns → Pull Habits

First, define Attention correctly:

Attention is sustained, voluntary, repeat engagement on channels a brand owns.

Attention is not opens. Not clicks. Not conversions. Attention is the relationship asset that makes all of those possible.

The Attention Stack

NeoMails Daily micro-engagements designed to be wanted — 60-second experiences that customers look forward to rather than ignore.

The shift:

• From campaign to ritual
• From promotional to valuable
• From episodic interruption to daily habit

Magnets Pull-based hooks that create the physics of attraction:

• Quizzes and predictions
• Games and challenges
• Utilities and tools
• Personalised recommendations

Social platforms understood this. Instagram embedded magnets. TikTok created loops. Wordle created rituals. Email was left magnetless — until now.

Mu (Attention Currency) Rewards engagement with tangible value. Every interaction earns points. Points accumulate into benefits.

This transforms attention economics from extraction (brand takes time, gives nothing) to exchange (brand gives value, earns attention).

Killing the Reacquisition Trap

ActionAds: When brands build engaged NeoMail audiences, that attention becomes valuable to non-competing advertisers.

• Advertising revenue flows to the brand, not to Google or Meta
• The programme becomes cost-neutral or profitable
• Retention funds itself

NeoNet: Cooperative, deterministic recovery across non-competing brands.

The logic:

• Customer dormant for Brand A but active for Brand B
• Brand A reaches them through Brand B’s NeoMails
• No auction. No bidding war. No adtech intermediation.

This is deterministic recovery: reaching a known customer through a known channel at a known cost.

Martech burns attention. NeoMarketing compounds it.

3

Synthesis

The Flywheel Reversed

The Old Vicious Cycle

Martech fails at retention → Customers lapse silently → Transitions go unmeasured → Only path back is ads → AdWaste grows → Martech gets blamed but economics don’t change → More customers lapse → Cycle accelerates…

The New Virtuous Cycle

Alpha aligns incentives → Agentic retains Best → Attention recovers Rest/Test → Customers stay → No reacquisition needed → Profits compound → Investment in retention grows → Cycle accelerates…

How The Three A’s Reinforce Each Other

• Alpha ensures vendors are incentivised to make Agentic and Attention work
• Agentic keeps Best customers from ever becoming Rest
• Attention catches Rest before they become Test — and recovers Test before they become adtech’s customers

Together, they shrink the customer pool that adtech can monetise.

The result: AdWaste shrinks. Profits grow. The $500 billion problem becomes the $500 billion opportunity.

The New Mantra

The Contrast

Traditional Martech: Lose customers. Pay twice. Repeat.

NeoMarketing: Never Lose Customers. Never Pay Twice. Pay Once. Profit Forever.

The Choice

Brands now face a binary decision:

• Continue paying the AdWaste tax ($500 billion annually)
• Adopt an operating system where retention is the default

Root Cause Resolution Map

Root Cause	The Failure	NeoMarketing Pillar	The Solution
#1: Input Pricing	Vendors paid for activity (messages, MTUs), not outcomes. No skin in the game. Failure was profitable.	ALPHA	Outcome-based economics (Beta + Alpha + Carry). Vendors profit only when brands profit. Retention becomes economically necessary.
#2: Complexity Gap	Platforms bloated with features (60-70% unused). No services ecosystem. Teams couldn’t operate the tools they bought.	AGENTIC	Marketing Agents (Insights, Segment, Content, Shopping) operate the complexity humans cannot. MGEs bridge until agents mature.
#3: Personalisation Mirage	Promised N=1, delivered stereotypes. 5-20 static segments for millions. “Not for me” messages trained customers to ignore.	AGENTIC	BrandTwins enable true N=1. Individual AI models for each customer, built in Twin Factory on ArtificialPeople foundations.
#4: Attention Blind Spot	Dashboards tracked campaigns, not relationships. Attention decay unmeasured. Transitions (Best→Rest→Test) invisible.	AGENTIC	Live Ledger tracks real-time P&L for every customer. Drift detected before dormancy. Intervention at optimal moments.
#5: Inbox Failure	Email trained to be ignored. Batch-and-blast as SOP. Push always decays, but martech pushed harder. 80% quarterly churn.	ATTENTION	NeoMails transform inbox from push to pull. Magnets create attraction. Mu rewards engagement. Daily habits replace quarterly campaigns.
#6: Best Customer Bias	Martech focused on 20% Best (easy wins). 80% Rest/Test abandoned. Missing middle handed to adtech.	ATTENTION	NeoMails and NeoNet focus on Rest/Test recovery. ActionAds fund the investment. No customer abandoned to drift.
#7: Adtech’s Advantage	Acquisition was easy (Agency→Budget→Clicks). Retention was complex. 90:10 budget split. ROAS culture infected retention metrics.	ALPHA	Alpha makes retention investment attractive. Vendors bring resources. Risk transfers from brand to partner. Retention properly funded.
#8: Reacquisition Mirage	50-70% of “acquisition” was reacquisition. Brands bid in auctions for customers they already owned. Double taxation on every lapsed customer.	ATTENTION	NeoNet enables cooperative, deterministic recovery. No auctions. No bidding wars. Known customers reached through known channels at known costs. Never pay twice.

 

**

The Inevitability

The question is not whether this change will happen.

The structural pressures are too great. AdWaste is too expensive. The AI capabilities are too powerful. The alternative — continuing to optimise a system designed to waste $500 billion annually — is not sustainable.

The question is who will lead.

• Will it be incumbent martech vendors, disrupting themselves?
• Will it be new entrants, building NeoMarketing natively?
• Will it be brands themselves, demanding outcome alignment?

The Final Word

The diagnosis is complete. The prescription is clear.

AdWaste is optional. The cure exists.

The only question is: who moves first?

Published January 31, 2026

1

Preamble, Root Cause 1

The marketing industry wastes $500 billion every year reacquiring customers that brands already own.

This is not a rounding error. It is not inefficiency at the margins. It is not the cost of doing business. It is the dominant use of marketing budgets: nearly 70% of advertising spend flows to reacquisition, retargeting, and “win-back” campaigns — money spent bringing back customers who should never have been lost in the first place.

The instinct is to blame the usual suspects. Google and Meta have built trillion-dollar empires on advertising. Surely they are the villains in this story?

They are not.

The uncomfortable truth is that AdWaste was not created by adtech. It was created by the industry that was supposed to prevent it: martech.

Martech — the sprawling ecosystem of CRM platforms, customer engagement tools, marketing automation, CDPs, and retention solutions — was built on a singular promise: help brands build better customer relationships. Keep customers engaged. Prevent churn. Make reacquisition unnecessary.

It failed.

Not because the tools lacked features. Not because marketers lacked ambition. But because the entire system — its economics, its architecture, its incentive structures, its measurement frameworks — was designed in ways that made failure not just possible, but inevitable.

This essay is a diagnosis, not an indictment. It is a systems-level post-mortem, not a blame exercise. The goal is not to attack vendors or marketers, but to understand how an industry built to solve retention became the upstream feeder system for the very platforms it was meant to replace.

AdWaste was not an accident. It was emergent behaviour — the logical output of eight structural failures that reinforced each other over two decades until they became inescapable.

These are the eight root causes of the $500 billion AdWaste crisis.

**

ROOT CAUSE #1: Input-Based Economics with Zero Accountability

(The Original Sin)

The foundational flaw: Martech priced activity, not outcomes.

What martech charged for:

• Messages sent
• Contacts/records stored
• MTUs / MAUs / events / API calls

What this created:

• Vendors got paid whether customers engaged or ignored
• Vendors got paid whether retention improved or collapsed
• No economic consequence for attention decay, silent churn, or poor lifecycle design
• Sending more was always rewarded — even when it accelerated fatigue

The perverse incentive:

• Failure at retention became profitable
• Success at retention became optional
• Customer loss became an externality — pushed downstream to adtech

The contrast: Adtech, despite its flaws, aligned with the marketer’s desperation for growth (CPA, CPC, ROAS). Adtech had skin in the game; martech just collected rent.

Evidence: Gartner data shows martech utilisation dropped from 58% (2020) → 42% (2022) → 33% (2024) — yet vendors kept getting paid the same.

Net effect: Martech had no skin in the retention game. Customer loss became someone else’s problem — eventually adtech’s opportunity.

Key line: “Martech sold activity, not outcomes. The meter ran whether customers stayed or left.”

2

Root Causes 2-4

ROOT CAUSE #2: Complexity Without a Services Economy

(The Ferrari-with-No-Driver Problem)

The gap: Martech platforms became extremely powerful — and extremely unusable.

What happened:

• Platforms bloated with features: journeys, rules, triggers, CDPs, AI engines
• 60-70% of features went unused across the industry
• CRM/lifecycle teams stayed small, under-skilled, lacking operational support

The missing layer:

• Unlike ERP (SAP, Oracle), martech never built a large services ecosystem
• SAP and Oracle spawned massive implementation partners (Accenture, TCS, Infosys, Deloitte) because contract values justified it
• Mid-market martech pricing couldn’t sustain integrators or operators — low entry costs meant no margin for services
• Vendors sold “self-serve sophistication” that wasn’t actually self-serve

The skills crisis:

• 64% of organisations acknowledge significant lack of internal martech/data/marketing operations expertise
• Most martech teams can’t distinguish between workable and harmful complexity
• The org chart disparity: five people optimising Facebook ads, one person managing 500,000 email subscribers

The behavioural outcome: When retention tools required armies and expertise, and acquisition required only a credit card, the credit card won.

Net effect: Sophistication existed on slides, not in execution. Platforms were reduced to broadcast tools. Acquisition platforms won by default because they were easier to operate.

Key line: “Martech sold Ferraris to people who never learned to drive. Then wondered why they took taxis instead.”

**

ROOT CAUSE #3: The Personalisation Mirage

(Promised N=1, Delivered Stereotypes)

The promise: One-to-one marketing at scale. The “Segment of One.”

The reality:

• 5-20 static segments for millions of customers
• Batch-and-blast as standard operating procedure
• “Women 25-34” or “loyal high-spenders” isn’t personalisation — it’s stereotyping at scale

Why true personalisation failed:

• Content creation didn’t scale (expensive, slow, operationally painful)
• Data files stayed thin — identity files (email, phone) rather than intent files
• No incentive to thicken customer understanding
• No services revenue to fund the work

The “Not for Me” problem:

• Generic messages trained customers to ignore, skim, disengage quietly
• Relevance collapsed fastest for Rest and Test customers — the 80% who needed personalisation most
• Post-purchase engagement especially weak: ecommerce → “One and Done” epidemic; BFSI → broken lead nurturing

The BRTN reality: 20% Best get campaigns (they’d buy anyway), 80% Rest and Test get ignored (the ones who needed personalisation most).

Net effect: Martech solved message delivery, not meaning. Customers disengaged quietly — quarter after quarter.

Key line: “Martech promised to know each customer. Instead, it sorted millions into a dozen buckets and called it personalisation.”

**

ROOT CAUSE #4: The Attention Blind Spot

(What Martech Never Measured)

The measurement failure: Martech built dashboards for the wrong things.

What martech optimised for:

• Opens
• Clicks
• Conversions
• Campaign ROAS

What martech did NOT track:

• Attention decay over time
• Engagement half-life
• Transition moments (Best → Rest → Test)
• Click Retention Rate (who keeps clicking quarter over quarter)
• Silent disengagement signals

The invisibility problem:

• No dashboards for attention decay
• No alerts for disengagement patterns
• No ownership of customer transitions
• Customers only became “visible” again when they were already lost — and being reacquired via ads

The timing failure:

• By the time a customer appeared in a “win-back” segment, they’d already completed the journey to dormancy
• The transition moments (Best→Rest, Rest→Test) — when intervention would be most effective and least costly — went unmonitored and unmanaged

Net effect: The most destructive failure mode — attention loss — remained invisible by design. Martech measured conversion, not continuity.

Key line: “Martech built dashboards for campaigns. It forgot to build dashboards for relationships.”

3

Root Causes 5-7

ROOT CAUSE #5: The Inbox Failure

(Squandering the Only Channel Brands Fully Owned)

The tragedy: Email should have been martech’s fortress. It became its greatest liability.

What email was:

• Fully owned (no platform dependency)
• Identity-rich (authenticated audience)
• Low marginal cost (near-zero per message)
• Universal (everyone has an email address)
• Habitual by nature (checked daily)

What martech did to it:

• Treated it as a static delivery pipe, not a relationship channel
• Made batch-and-blast the standard operating procedure
• No innovation around habit formation, pull, or daily value
• No answer to attention half-life, habituation, or inbox fatigue

The physics violation:

• Push always decays, but martech kept pushing harder
• When engagement dropped, response was: more frequency, louder subject lines, deeper discounts
• Expansion was the answer to every problem, never examination

The contrast: Social platforms understood pull mechanics. Instagram embedded magnets — micro-moments of reward, validation, curiosity. TikTok created daily habits. Wordle created rituals. Email was left magnetless.

The result:

• 80% quarterly attention churn (4 of 5 clickers disappear within 90 days)
• Not by unsubscribing, but by quietly stopping
• 90%+ of emails became ignorable by design
• How can an industry build itself when 90% of its messages are ignored?

Net effect: Email didn’t die. It was trained to be ignored. Owned attention eroded predictably.

Key line: “Email didn’t die. Martech killed it — one ignored message at a time.”

**

ROOT CAUSE #6: The Best Customer Bias

(Fishing Where the Fish Already Are)

The blind spot: Martech focused on the extremes and abandoned the customers who determined profitability.

What martech focused on:

• Best customers (high-value, high-frequency)
• High-intent moments (cart abandonment, browse abandonment)
• Customers who would buy anyway

Why:

• They convert easily
• They make dashboards look good
• It’s the marketing equivalent of fishing in a stocked pond — impressive catches, but not sustainable growth

The “Missing Middle” abandoned:

• Rest customers (showing early disengagement) got the same generic messages as everyone else
• Test customers (90+ days dormant) deemed lost and handed to adtech for reacquisition
• 70-80% of customer bases received almost no strategic attention

The typical brand reality:

• 100,000 customers: 20,000 Best keep buying, 40,000 Rest slowly disengage, 40,000 Test already gone
• Martech focuses on the 20,000
• Adtech profits from the 80,000

Net effect: The customers who needed martech most were abandoned earliest — and handed directly to adtech. A perfect parasitic relationship: martech’s neglect became adtech’s nourishment.

Key line: “Martech focused on the 20% who would buy anyway, while the 80% who needed nurturing silently walked away.”

**

ROOT CAUSE #7: Adtech’s Asymmetric Advantage

(The Easy Button Won)

The divergence: While martech stagnated in complexity, adtech compounded in simplicity.

What adtech perfected:

• The ABC model: Agency → Budget → Clicks
• Call an agency, allocate budget, receive traffic — seductively simple
• Predictable, measurable, immediate results
• Minimal expertise required

What adtech created:

• A slew of agencies industrialised acquisition
• The marketer’s job of acquisition became easy, measurable, career-safe
• Performance marketing created ROAS culture: worship immediate results

The ROAS infection:

• This thinking spread to retention channels — relationship-building got measured like acquisition campaigns
• Even email and CRM got judged by short-term conversion, last-click attribution
• The tyranny of the urgent over the important

The resource disparity:

• Budget allocation: 90% acquisition, 10% retention — nine times more spent on the expensive activity
• Team allocation: adtech teams are specialised and well-resourced; martech teams are “the email person”
• Technology investment: 10:1 in favour of adtech ($200K-500K vs $20K-50K for mid-market)
• Retention became “email — it’s basically free”

The dependency trap: Because martech wasn’t holding customers (high churn), brands had to constantly refill the “leaky bucket” via adtech.

Net effect: Growth became a budget function, not a relationship function. Attention was outsourced instead of built. Adtech filled the vacuum martech created.

Key line: “When retention required expertise and acquisition required only a credit card, the credit card won.”

4

Root Cause 8, Synthesis

ROOT CAUSE #8: Reacquisition Disguised as Growth

(The AdWaste Flywheel)

The final indignity: Most “acquisition” spend is actually reacquisition in disguise.

The silent churn reality:

• 80% of engaged customers drift quarter over quarter
• Drift is invisible (no dashboards, no alerts — see Root Cause #4)
• Customers don’t complain, don’t unsubscribe — they just stop

What brands do to maintain revenue:

• Spend more on acquisition
• Buy traffic, retarget broadly, bid in auctions

The reacquisition mirage:

• 50-70% of “acquisition” spend targets people already in brand databases
• Brands are bidding in auctions to reach customers they already “own” but have tuned out
• These customers disengaged due to Root Causes #3-6
• Attribution models break; budgets balloon; no one questions why the “new” looks suspiciously like the old

The double taxation:

• First, brands lose the customer’s lifetime value — all future purchases vanish
• Then brands pay again — often more than original CAC — to bring them back through ads
• If they don’t, competitors will
• Either way, profit vanishes into a permanent P&L leak

The economics:

• Every abandoned customer is worth their weight in gold to adtech platforms
• Google and Meta happily sell them back at 5-7x the cost of retention
• Martech’s retention failure directly subsidises adtech’s record profits

Net effect: AdWaste is not inefficiency. It is not bad execution. It is the logical output of the system. Brands pay platforms to reach customers they already own, because they trained those customers to ignore the channels they control.

Key line: “The ultimate absurdity: brands pay twice for customers they already own — because martech trained those customers to stop listening.”

**

SYNTHESIS: The System That Could Only Produce AdWaste

These eight failures weren’t isolated — they formed a self-reinforcing system:

1. Input pricing → No accountability → Features bloated without discipline
2. Complexity without services → Low utilisation → Defaults to batch-and-blast
3. Personalisation failure → “Not for me” messages → Quiet disengagement
4. Attention blind spot → Decay unmeasured → Transitions invisible
5. Inbox abuse → Owned channels collapse → Email trained to be ignored
6. Best-customer bias → 80% abandoned → Rest and Test hand-delivered to adtech
7. Adtech ease → Budget addiction → Martech starved of investment and talent
8. Reacquisition mirage → Auction dependency → AdWaste becomes structural

The flywheel of failure: Martech fails at retention → Customers lapse silently → Transitions unmeasured → Only path back is ads → AdWaste grows → Martech gets blamed but economics don’t change → More customers lapse → Cycle accelerates…

The result:

• 80% quarterly attention churn
• 90% of marketing budgets flow to adtech
• 70% of that is reacquisition (buying back known customers)
• $500 billion annual AdWaste globally

The uncomfortable truth: AdWaste isn’t a bug in martech. It’s a feature of its economics. Martech didn’t lose to adtech — it became adtech’s most valuable upstream partner.

**

The Question That Remains

Martech wasn’t beaten by adtech. It was abandoned by design:

• Abandoned by pricing that rewarded activity over outcomes
• Abandoned by complexity that exceeded human capacity to operate
• Abandoned by personalisation promises that collapsed into stereotypes
• Abandoned by measurement frameworks that couldn’t see attention decay
• Abandoned by channel abuse that destroyed owned attention
• Abandoned by strategic blindness that ignored 80% of customers
• Abandoned by ease of alternatives that required no expertise
• Abandoned by attribution that couldn’t distinguish acquisition from reacquisition

The $500 billion AdWaste crisis isn’t a market inefficiency. It’s a system working exactly as designed — just not for brands.

The question now isn’t “who’s to blame?”

It’s “what must fundamentally change?”

That’s the subject of the next essay.

Published January 30, 2026

1

Commentary

A few days ago (Jan 25), Wall Street Journal had a story entitled “Wall Street Has Fallen Out of Love With Software Stocks” which began thus: “Software companies’ pitch to investors could use an upgrade. Once a favorite of Wall Street, software stocks have been sliding lately, with investors increasingly concerned about how the sector could be upended by their newest crush: artificial-intelligence companies. Rocked by the emergence of “vibe coding”—the practice of using AI tools to quickly produce apps and websites—software heavyweights Salesforce, Adobe and ServiceNow are all down at least 30% since the start of last year. An S&P index of small and midsize software stocks is also down more than 20% over that period, with declines accelerating this month after the introduction of Anthropic’s Claude Code, an AI tool that industry insiders have said can dramatically shrink the time it takes to build even complex software.” The story added: “In reality, few investors and analysts think that software companies will become obsolete in the foreseeable future. The more pressing risk is that it could become more difficult to increase revenue, as customers experiment with other options rather than paying more for the usual updates and add-ons, RBC’s [Rishi] Jaluria said.”

This was one in a series of such stories that I had come across in the past few weeks.

FT (Jan 21): ““Software is not at all about the code or about the technology. Software is about your domain knowledge,” said Thoma Bravo co-founder Orlando Bravo…The plunge in software valuations, driven by fear of an existential threat from artificial intelligence, is creating a “huge buying opportunity,” Orlando Bravo, the firm’s co-founder, told the FT [in Davos]… Bravo’s comments come after a plunge in software valuations in recent weeks. Software is one of the US stock market’s worst-performing sectors so far this year, with an index tracking the group down about 7 per cent over the past three weeks.”

Bloomberg (Jan 18): “All told, a group of software-as-a-service stocks tracked by Morgan Stanley is down 15% so far this year, following a drop of 11% in 2025. It’s the worst start to a year since 2022, according to data compiled by Bloomberg… The latest selloff has exacerbated an already yawning gap between the performance of software companies and other areas of the tech sector. Anxieties about competition from upstart AI services are overshadowing characteristics like hefty profit margins and recurring revenue that for years made the group attractive in the eyes of market pros.”

Sherwood (Jan 16): “The relentless slide in software stocks continues…The growing adoption of Claude Code, and more recently, the launch of Claude Cowork by Anthropic, has been an attention-grabbing moment as to the power of AI agents and how they can be housed and operated solely under one highly integrated user interface.”

Heise (Jan 22): “The sell-off can hardly be classified as more than an ordinary correction; the price losses since the beginning of the year alone seem more like they’re from a textbook for major stock market crashes. Salesforce, Adobe, and Oracle are all in double-digit losses in 2026. ServiceNow, Atlassian, and HubSpot are even losing 30, 40, 50 percent at times, as a price overview from Fiscal.ai on social media impressively illustrates. The narrative that has been playing out on Wall Street trading desks for months is: “AI eats software.”…If autonomous AI agents handle tasks in the future that used to require entire corporate departments, why pay for dozens of Salesforce or Workday licenses? If AI reviews, formulates, and closes contracts, what’s the need for widespread DocuSign subscriptions? And if generative image and design models deliver in seconds what teams of graphic designers used to do – how many Photoshop licenses does a company need from Adobe then? Investors’ fear: productivity explodes, but software providers’ revenues implode. AI is celebrated, but it could prove to be pure poison for the seat economy.”

CTech (Jan 25): “Software companies once beloved by investors, from Silicon Valley to Wall Street, are fading as talk of the “death of SaaS” gains momentum. Until recently, much of the threat seemed to come from a handful of trendy vibe-coding startups, platforms that allow non-programmers to build simple applications using verbal commands…As AI agents commoditize software, companies such as Wix, monday and Nice confront a forced reinvention of growth, pricing and value.”

Jason Lemkin: “Traditional B2B software and SaaS is under assault.  The leaders are all still growing, but in most cases slower than ever. Stock prices are under intense pressure in 2026 for anyone not growing > 20% or more.” He listed six threat vectors: fewer seats, AI budget shift, slow teams, product decay, TAM trap, price increases. He added: “The question your customers are asking themselves: “Do I want to use this product, or do I have to use this product?” If the answer is “have to,” you’re living on borrowed time.  Customers may stay because they are prisoners, or because you have their data.  But they may no longer want to.”

About the capital shift to AI, Jason Lemkin wrote: “2025 saw over $225 billion deployed into AI startups—46% of all venture capital. Five companies alone (OpenAI, Scale AI, Anthropic, Project Prometheus, and xAI) raised $84 billion. That’s 20% of all VC funding in one year going to five companies.” He added: “The market is saying: we’ll pay up for the infrastructure that makes AI possible—and we’ll make those employees generationally wealthy—but we’re deeply skeptical that incumbent software can adapt fast enough… Private AI companies are getting 3-10x the multiples of public SaaS incumbents. The market is pricing in wholesale disruption of existing software categories.”

Rohan Paul commented on a Satya Nadella interview: “”Satya Nadela is basically describing the death of the traditional SaaS model. Explains the AI agentic future, and where the “value” lives. Because business logic is moving from the software application to the AI agents. Currently, you buy software for its specific features and rules. Nadella argues that in the future, software apps will essentially become dumb databases (“CRUD”) or simple tools. The AI Agent will hold all the intelligence, orchestration, and reasoning, simply updating the databases as needed. The software becomes a commodity; the AI becomes the “brain” and the worker.”

**

My thesis

The commentary above captures what’s happening. But most analysis stops at diagnosis. I want to push toward prescription.

B2B software isn’t in a cyclical dip — it’s being structurally repriced. The subscription model that powered two decades of compounding (seats, renewals, expansion) is now being squeezed by multiple forces at once: AI-driven capability deflation, seat compression, vendor sprawl backlash, budget reallocation to AI and security, and the rise of faster, leaner AI-native competitors. Public market sentiment is reflecting this shift, with investors openly asking whether AI changes what software is “worth.”

The conventional fixes — add AI features, cut costs, buy competitors — are necessary but not sufficient. They defend existing revenue; they don’t create new profit pools. The winners will add new revenue engines that align with value created and attention captured — not only with headcount and licences.

This series examines the crisis, its root causes, the inadequacy of conventional responses, and a new revenue playbook built on two additional engines: performance fees (outcomes-based revenue) and attention yield (monetising engagement, not seats). The final part provides an implementation framework — the two-track approach — for building new revenue streams without destroying what’s working.

Sources & Influences

This series draws on public commentary and reporting including Jason Lemkin (SaaStr), Wall Street Journal, Financial Times, Bloomberg, Sherwood News, Heise, CTech, Gartner, Vertice, Zylo, Iconiq, and SemiAnalysis. Direct quotes are attributed in-text.

The interpretations, the “new revenue playbook” framework (performance fees + attention yield), and the two-track implementation model proposed in Parts 5–6 are my own, developed through work at Netcore and my writings.

As with my other writing, I combine my thinking with assistance from AI (Claude and ChatGPT).

2

The Crisis

The repricing signal: when recurring revenue stops feeling “inevitable”

For years, B2B software earned a premium because it looked like the closest thing public markets had to a perpetual motion machine: multi-year contracts, high gross margins, low churn, and expansion built on seat growth. That story hasn’t ended — but it has lost its certainty.

Over recent weeks and months, software has noticeably underperformed the broader tech indices. Several category-defining names — Salesforce, Adobe, ServiceNow, HubSpot, Atlassian, Workday, Intuit — have seen double-digit declines over short windows and meaningful drawdowns (30-50%+) from their peaks. The precise numbers will keep shifting, but the signal is stable: investors are pricing higher uncertainty into software’s future cash flows.

What’s different this time is not merely rates or rotation. It’s a narrative inversion: AI was supposed to be the accelerant for software. Instead, the market is treating AI as a potential solvent.

One quote captures the emotional turn perfectly: “The narrative has really shifted… Investors have gone from initially thinking that software companies could benefit from AI to asking, ‘Is AI just the death of software?'” — Rishi Jaluria, RBC Capital Markets [WSJ]

That single line is doing a lot of work. It doesn’t mean software disappears. It means the market is asking whether the traditional software bargain — “pay us forever for capability” — survives in a world where capability can be generated, copied, and embedded faster than ever.

The valuation collapse

The multiple compression has been brutal.

A basket of software-as-a-service stocks tracked by Morgan Stanley is now trading at roughly 18 times forward earnings — its cheapest level on record. The historical average over the past decade exceeded 55 times. That’s not a discount. That’s a fundamental repricing of what software companies are worth.

The reason software commanded lofty multiples was simple: subscription-based recurring revenue that you could extrapolate into the future almost forever. Customers got locked in. Switching costs were high. Growth was predictable.

“It is hard to know what multiple they should be trading at if they’re going up against AI agents that are running 24/7 and have the ability to complete tasks, with big projects getting done in a day.” — Bryan Wong, Osterweis Capital Management [Bloomberg]

The old certainties are gone. And so are the old valuations.

Meanwhile, AI-native companies — both foundation model providers and application-layer startups — are commanding multiples that make traditional SaaS look like deep value investing. Private AI companies are raising at 50-100x revenue while public software trades at 6-12x. The market is pricing in wholesale disruption.

The Claude Code moment

Fear becomes acute when a tool makes the abstract feel concrete.

Recent coverage around Anthropic’s Claude Code gave investors a vivid mental model: building software compresses from months to days for a widening set of use cases. Developers reported completing complex projects in a week that would have taken a year. Non-engineers built their first applications without ever learning to code.

“I cannot stress enough that Claude Code is the ChatGPT moment repeated. You must try it to understand… This is going to hurt a large part of the software industry.” — Doug O’Laughlin [SemiAnalysis]

You don’t need to believe every hyperbolic claim to accept the direction: the cost and time to produce good-enough software is falling sharply. That attacks the pricing umbrella for broad swathes of application software.

The divergence: public software vs. private AI

At the same time public software is getting hammered, private markets are pouring money into AI infrastructure and AI-native applications at valuations that imply massive future profit pools.

Foundation model companies have raised tens of billions at valuations that would have seemed absurd two years ago. Application-layer AI startups are reaching $100 million in annual recurring revenue in one to two years — versus five-plus years historically for traditional SaaS.

“We’re living through something we’ve never quite seen before in B2B software. It’s arguably the worst time in our history to be invested in public software stocks. And simultaneously, the best time in our history to be invested in private hot AI-fueled startup stocks.” — Jason Lemkin [SaaStr]

The exact valuations will change. The underlying divergence is what matters. Public markets are demanding proof. Private markets are funding possibility.

The core question

This is what the market is now forcing on every B2B software company:

If AI can increasingly do what your product does — or enable customers to assemble substitutes — what exactly are customers paying you for, and why will that payment scale?

Understanding why this is happening structurally — beyond “AI fear” — is essential.

My thesis in brief: If seats compress, software needs new revenue primitives — not just new features. I see two: outcome-tied fees (paying for measurable improvement, not capability) and attention yield (monetising engagement, not headcount). The rest of this series explains why conventional responses fall short, and how to build these new engines without destroying what’s working.

3

Root Causes

Six compounding pressures (only one is “AI competition”)

Jason Lemkin’s recent SaaStr analysis identified six threat vectors attacking traditional B2B software. I want to build on his framework — not just list the pressures, but explore why they compound, and what that compounding means for revenue strategy. The forces don’t operate independently; they reinforce each other in ways that make “add AI features” insufficient as a response.

The temptation is to treat the current sell-off as sentiment that will mean-revert. That’s dangerous. B2B software faces multiple converging pressures — only one is direct AI competition. The others are economic and organisational shifts that would be squeezing the sector even if ChatGPT had never launched.

My lens: from threat vectors to revenue architecture

The threat vectors are real. But diagnosing pressures isn’t the same as prescribing solutions. Most commentary stops at “software is under pressure” or “add AI features.” I want to push further: if the unit economics of seats are breaking down, what new units of value can software companies sell?

Three shifts matter most for revenue strategy:

• The CFO reversion. Enterprise spend is shifting from “experiments” to “guarantees.” Risk transfer becomes the product. Outcome-based pricing isn’t just attractive — it’s what procurement increasingly demands. CFOs are tired of paying for capability without accountability.
• The orchestration migration. As Satya Nadella suggests, apps may become systems of record while AI agents handle orchestration and reasoning. If that’s true, value migrates to measurement, guarantees, and the data layer — not workflows and UIs.
• The attention opportunity. B2B software generates enormous engagement that’s never been monetised. Email platforms see billions of opens daily. Collaboration tools capture hours of attention. That’s a latent revenue pool waiting to be unlocked — if companies can move beyond the “ads are beneath us” mindset.

With that lens, let’s examine the six pressures — and what each implies for revenue architecture.

1. Seat growth is slowing (and AI amplifies the slowdown)

Per-user pricing won’t vanish overnight — even AI-native darlings like Cursor and Anthropic charge per seat. But the seat-growth engine is no longer the dependable escalator it was. The unit of pricing may survive; the expansion dynamic has stalled.

Companies from Workday on down are seeing customers commit to lower headcount on renewals. As Workday CEO Carl Eschenbach acknowledged [SaaStr]: “We are seeing customers committing to lower headcount levels on renewals compared to what we had expected. We expect these dynamics to persist in the near term.”

Some of this is simply headcount slowdown. Tech hiring has flattened. Some companies are holding headcount flat for years while still growing revenue significantly — proof that productivity gains are real and that the old “more employees = more seats” equation is breaking.

AI agents accelerate the compression. The relationship isn’t linear or predictable, but every autonomous agent handling work previously done by humans creates downward pressure on seat counts. The maths works against traditional expansion models.

The data tells the story. Net revenue retention at leading SaaS companies has flattened or declined. Some companies are seeing enterprise customer counts actually decline while still generating strong free cash flow — the classic signs of harvest mode. Revenue holds; growth stalls.

Revenue implication: If seats don’t expand, you need revenue that scales with transactions or outcomes, not headcount. This is the core case for performance-based pricing.

In martech specifically, the seat question is even more acute: marketing teams are shrinking while marketing automation demands grow. The gap is being filled by agents, not hires. A platform priced per marketer faces structural headwinds; a platform priced per incremental sale does not.

2. SaaS sprawl backlash and “SaaS inflation”

Buyers are drowning in tools — and increasingly aware that many of those tools creep up in price annually.

SaaS prices have been rising at roughly four to five times general inflation. The average enterprise now spends significantly more per employee on SaaS than just two years ago. Analysis suggests that a substantial majority of some vendors’ recent growth came from price increases, not new customers.

As Gartner’s John-David Lovelock observed [SaaStr]: “The cost of software is going up and both the cost of features and functionality is going up as well thanks to GenAI.”

This creates a vicious cycle. Price increases eat incremental budget. Reduced budget means less room for new purchases. Less expansion pressures vendors to raise prices again. The spiral continues until something breaks.

Most IT budget uplifts are being absorbed by renewals, security posture, and foundational AI commitments. There’s less discretionary room for “yet another tool” — and even less appetite for price increases justified by AI features users didn’t ask for.

Revenue implication: The procurement trap is real. Vendor consolidation pressure makes “incremental modules” harder to sell than “revenue engines.” If you’re asking for more budget, you need to show you’re generating budget — through measurable outcomes or cost offsets. Pure capability expansion hits a ceiling.

3. The AI + security budget gravity well

Whether you love it or hate it, a growing share of incremental IT spend is being pulled into foundation models, AI tooling, and security posture.

Enterprise leaders expect dramatic growth in LLM budgets over the next year. AI has graduated from experiment to core operating expense. The majority of AI use cases are now purchased rather than built internally.

Foundation model companies alone are generating tens of billions in combined annualised revenue — consuming a material share of all incremental IT budget.

As Jason Lemkin [SaaStr] put it bluntly: “If you’re not tapping into AI budget, you’re fighting for scraps.”

If you can’t articulate how you replace humans, dramatically augment humans, or enable the previously impossible — you’re competing for a shrinking pool of non-AI budget. The “incremental productivity improvement” pitch that worked for a decade now sounds quaint next to “we eliminated three roles.”

Revenue implication: Outcome-based pricing taps into AI budget naturally, because outcomes are what AI budget is for. A platform that guarantees measurable improvement competes for AI budget. A platform that offers “AI-powered features” without measurable impact competes for scraps.

4. The efficiency gap: AI-native companies are faster and leaner

AI-native companies are showing startling revenue-per-employee figures and shipping velocity. They’re built around new production functions: smaller teams, higher leverage, faster iteration loops.

The efficiency gap is becoming a chasm. AI-native startups are averaging five to six times the revenue per employee of traditional mature SaaS. They’re reaching $100 million ARR in one to two years versus five-plus years historically.

When you can achieve the same scale with dramatically less capital and fewer people, you can move faster, experiment more, and capture market share while competitors are still in planning meetings.

The key point isn’t the precise number — it’s the magnitude of the gap. And it’s not just about headcount efficiency — it’s about iteration velocity. AI-native teams ship daily because they don’t have organisational antibodies resisting change. Traditional teams are still debating what to build while competitors are already measuring what works.

Revenue implication: Speed compounds. Companies that can prove value in 90 days will win deals over companies that need 12-month implementations. Outcome-based models force this discipline: you can’t charge for outcomes you haven’t delivered, so you’re incentivised to deliver fast. The pricing model becomes a forcing function for operational excellence.

5. TAM traps and maturity

Some categories are simply maturing.

When expansion turns into harvesting — price per seat, margin maximisation — markets stop paying for “forever growth.” When management talks about “seat expansion” and “price per seat increases” instead of customer acquisition, you’re watching a company hit its TAM ceiling.

The warning signs are clear across the industry: revenue up, customers flat or down. Growth decelerating from 30%+ to single digits. Entire strategies centred on extracting more from existing customers rather than acquiring new ones.

Markets don’t fear collapse. They fear decades of mediocrity — the slow fade of a company that’s stopped growing but hasn’t yet died.

Revenue implication: TAM traps are category-specific, but the escape route is universal: find new units of value. If you’ve saturated seat expansion in your category, you need revenue streams that don’t depend on seats. Performance fees and attention yield are TAM-agnostic — they scale with customer success and engagement, not with headcount in a saturating category.

6. The product experience gap

This is the most uncomfortable pressure to acknowledge.

AI-native products frequently feel magical in a way legacy SaaS does not. From Claude to the best AI-native applications, these products don’t just improve productivity. They create delight. Instantly.

The experience gap isn’t subjective. Chatting with data beats navigating via clicks. Natural language beats form fields. Getting an answer in seconds beats clicking through five screens. The interface paradigm has shifted, and products designed for the click-and-navigate era feel like using a fax machine.

When users can build functional apps in minutes or get meaningful output in seconds, the bar for what constitutes a “great product” has permanently shifted.

Revenue implication: Experience drives engagement. Engagement drives attention. Attention is monetisable. The companies that create delightful, habitual experiences have an asset they’re not exploiting: the attention surface. Meanwhile, companies with clunky experiences have neither the engagement to monetise nor the goodwill to raise prices.

**

The meta-shift: where intelligence lives

Beneath all six pressures lies a deeper structural transformation.

Business logic is moving from the software application to AI agents. Today, you buy software for its specific features and rules. The application holds the intelligence. Tomorrow, software applications may become systems of record and CRUD (create, read, update, delete) layers — simple tools for storing and retrieving information — while AI agents hold all the intelligence, orchestration, and reasoning.

As Doug O’Laughlin wrote: “I believe that the future role of software will not have much ‘information processing’, i.e., analysis. Claude Code or Agent-Next will be doing the information synthesis, the GUI, and the workflow. That will be ephemeral and generated for the use at hand…Most SaaS companies today need to shift their business models to more closely resemble API-based models to align with the memory hierarchy of the future of software. Data’s safekeeping and longer-term storage are largely the role of software companies now, and they must learn to look much more like infrastructure software to be consumed by AI Agents. I believe that is what’s next.”

What becomes worthless in this world: faster workflows, better UIs, smoother integrations. All of that can be generated on demand.

What becomes valuable: persistent information, APIs, proprietary data — and the ability to prove outcomes and own attention. If the intelligence layer commoditises, the measurement layer and the engagement layer become the new sources of defensible value.

**

The compounding problem

The important conclusion: these forces compound.

Seat compression reduces expansion. Price hikes create backlash. AI budgets siphon incremental spend. AI-native speed widens competitive gaps. Mature categories saturate. Experience gaps accelerate switching intent. Each weakness exposes the others. Each pressure intensifies the rest.

A company facing one of these pressures can adapt. A company facing all six simultaneously — which is most of B2B software — needs more than incremental responses.

If that’s true, “add AI features” is necessary — but not sufficient. The question becomes: what is the industry doing about it.

4

What the Market Is Trying — And Why It Falls Short

Defence is not a revenue strategy

The industry is not asleep. Executives are responding. Boards are asking hard questions. Capital is being reallocated.

But most responses are defensive: they protect today’s revenue model rather than create tomorrow’s.

Response 1: “Add AI features”

The most common response is to bolt AI capabilities onto existing products. Copilots. Assistants. Agent add-ons. Generative features. Every enterprise software company now has an “AI strategy.”

The problem: adding AI doesn’t automatically change willingness-to-pay — especially if buyers believe AI features will commoditise and converge.

Even investors are noticing gaps between “AI messaging” and “AI monetisation.” Major vendors have touted AI adoption, but it hasn’t moved the revenue needle significantly. Some didn’t update AI-related measures in their earnings reports — a telling omission.

Adding features that customers expect but won’t pay more for is a treadmill, not a strategy. If AI features become table stakes — something every competitor offers — they don’t justify premium pricing.

Response 2: “Domain knowledge is our moat”

This is the Thoma Bravo argument — and it’s partly right. “Software is not at all about the code or about the technology. Software is about your domain knowledge. Most software companies know a specific vertical, a specific process, a specific function so well that there are three to five companies in the world that know it, and about 20 individuals in the world that really, really know it. That is the franchise. That is the value. That is what you cannot replicate.”

Orlando Bravo is right that payroll, compliance-heavy workflows, and regulated systems don’t evaporate overnight. Domain expertise is genuinely hard to replicate. Integration moats are real.

But this is primarily a defence of durability, not a plan for new revenue expansion under seat compression. Domain knowledge protects existing revenue; it doesn’t create new revenue streams. And foundation models are building domain knowledge faster than many expected.

Bravo himself acknowledged that companies without specialisation are “absolutely vulnerable.” The question is whether even specialised companies can grow when their core revenue model is under pressure.

Protecting the franchise isn’t the same as expanding it.

Response 3: Cost cutting and “efficient growth”

Faced with slowing growth, many software companies have turned to the efficiency playbook. Layoffs. Margin expansion. “Disciplined execution.”

Some companies are generating strong free cash flow even as revenue growth stalls. Management talks about operational efficiency and 50% margins.

This is harvest mode. It works — for a while. Cash flow remains strong. Investors focused on profitability may appreciate the discipline.

But cost-cutting is an anaesthetic, not a cure. You can’t cut your way to a new growth curve. Eventually, there’s nothing left to cut. And if your category’s pricing power is under pressure, efficiency alone becomes a slow glide path to irrelevance.

Response 4: Consolidation / M&A

Private equity sees opportunity in repriced assets. Thoma Bravo has raised tens of billions for software deals, calling the sell-off a “huge buying opportunity.”

The thesis is straightforward: buy beaten-down software companies at discounted valuations, optimise operations, extract cash flow, and eventually exit. Consolidation reduces competition and creates scale.

But consolidation isn’t innovation. It’s the same revenue model with fewer players. The PE playbook — optimise margins, reduce costs, raise prices — doesn’t solve the structural challenges. It just extends the harvest phase.

And even PE isn’t immune. Some Thoma Bravo deals have reportedly soured due to AI-related issues. Other major PE firms have cut software exposure or even shorted software debt over AI fears.

When the smartest money in the room is hedging its software bets, consolidation alone isn’t the answer.

Response 5: “Wait for the narrative to turn”

Some bulls argue that the sell-off is overdone. AI will ultimately be a tailwind for software, not a headwind. The total addressable market will expand. Incumbent advantages in distribution and data will prevail.

They may be right — eventually. But hope is not a strategy.

Even if public software rebounds, the underlying structural forces remain: agentic workflows, budget shifts, buyer fatigue, speed gaps. The structural changes don’t reverse on their own. Even if AI expands the total addressable market, there’s no guarantee incumbents will capture the expansion.

Response 6: “Foundation models will struggle to build business software”

This is the most sophisticated defence, articulated by Orlando Bravo in a Financial Times op-ed.

The argument: OpenAI and Anthropic face the same challenge as every tech giant before them. Building enterprise software from scratch requires decades of industry knowledge, thousands of integrations, deep understanding of regulations, and the trust of large enterprises. Code is easy. Domain knowledge is hard.

As Orlando Bravo said, “Companies like OpenAI trying to build business software face the same challenge as every tech giant before them: creating entire business systems from scratch. What’s hard is the decades of industry knowledge, thousands of existing connections to other software, deep understanding of industry-specific regulations and the built-up trust of large enterprises.”

This is reasonable. Foundation model companies would struggle to replicate what incumbents have built over decades.

But the argument assumes the competition is foundation models building business software from scratch. It ignores the possibility of foundation models partnering with nimbler players, or enabling customers to build their own solutions, or simply commoditising the intelligence layer while incumbents are stuck with legacy architectures.

More importantly, it’s a defensive posture. “They can’t beat us” is different from “here’s how we win.”

What’s missing

Notice what all these responses have in common: they focus on defending existing subscription revenue.

Almost everything starts with: “How do we defend subscriptions?”

The better question is: What new revenue engines can a software company add that don’t depend on seat expansion and don’t require endless price hikes?

5

The New Revenue Playbook

Beyond subscriptions: two additional revenue engines

If subscriptions are under assault, what comes next?

Not “better subscriptions.” Not subscriptions with AI features. Entirely new revenue streams that align payment with value delivered.

This is the heart of the thesis — and it’s strongest when framed as additive, not as a wholesale replacement for subscriptions.

Think of the modern B2B software company as needing three revenue modes, of which two are new:

1. Subscription (stability — the foundation that funds everything)
2. Performance / outcomes (upside aligned to value created)
3. Attention / ecosystem yield (monetise engagement and distribution, not headcount)

Why these new engines?

Both “new engines” share four properties that the market is now rewarding:

Property	What it means
Variable	Scales with customer success or engagement
Aligned	Paid when value is realised
Non-linear	Can grow without seats
Defensible	Requires measurement, data, network, or workflow position

These properties escape the forces crushing traditional subscriptions. They don’t depend on headcount growth. They don’t require price increases that alienate customers. They align vendor and customer incentives. And they create moats that generic AI models can’t easily replicate.

Engine #1: Performance fees (outcomes-based revenue)

Principle: Stop charging only for capability. Charge for measurable improvement.

This isn’t new as an idea — consultancies and hedge funds have done it for decades. But software has historically avoided it because subscriptions are easier to forecast.

The point now is: forecastability is being repriced anyway. And alignment is becoming a competitive advantage.

The structure (borrowing from finance)

The terminology comes from hedge fund economics, but the application is direct:

• Beta is the baseline — what would have happened anyway. In finance, beta is market return; in outcome-based software pricing, beta is the customer’s performance without your intervention. This isn’t a fee; it’s the benchmark against which value is measured.
• Alpha is the outperformance — the incremental improvement above beta. The vendor charges a percentage of this uplift. No uplift, no payment. This is pure alignment: you only earn when you create measurable value.
• Carry is the long-tail participation — if Alpha persists over time, the vendor continues to share in the sustained improvement. This rewards durable impact, not one-time spikes.

A practical performance fee structure looks like this:

Component	What it means
Beta (Baseline)	The counterfactual — what performance would have been without intervention. This is the measurement benchmark, not a fee.
Alpha (Performance fee)	% of incremental value delivered above Beta. The vendor only earns when they create measurable uplift.
Carry (Long-tail share)	If Alpha persists over time, the vendor continues to participate in sustained improvement.

This does three useful things simultaneously:

• Makes the CFO feel safer. “Pay when it works” removes risk from the customer’s perspective. No uplift means minimal payment. This is a compelling proposition for finance leaders tired of paying for capabilities they’re not sure deliver value.
• Forces measurement rigour. Outcome pricing requires instrumentation, data trust, customer governance, and a shared definition of “incremental.” This rigour becomes a competitive advantage — and a moat.
• Escapes seat compression. Revenue is tied to transactions or outcomes, not headcount. If AI reduces the humans involved, the vendor still gets paid on results.

I’ve explored this extensively in my NeoMarketing essays, where the application to martech is direct: performance fees linked to incremental sales generated, with measurement built into the channel infrastructure. But the model generalises. Any B2B software category with measurable customer outcomes can adopt this structure.

Where it applies (beyond martech)

Performance fee models can work across B2B software wherever outcomes are measurable:

Vertical	Traditional Model	Performance Fee Model
Sales Tech	Per-seat CRM	% of pipeline conversion uplift
HR Tech	Per-employee platform fee	% of retention improvement or hiring cost reduction
Fintech (B2B)	Platform fee	% of fraud prevented or collections improved
Supply Chain	Per-user licence	% of cost savings or delivery improvement
Customer Support	Per-agent pricing	% of containment rate improvement or CSAT uplift
DevOps	Per-seat IDE	% of deployment velocity improvement or incident reduction

The common thread: tie revenue to outcomes that customers actually care about, not capabilities they may or may not use.

The key requirement

Outcome pricing is not a pricing page change — it’s an operating model.

It requires:

• Instrumentation (measuring what matters)
• Data trust (agreeing on the source of truth)
• Customer governance (pre-agreed methodology)
• Attribution clarity (defining what’s “incremental”)
• Control groups or credible quasi-experiments (proving causation)

This is hard. That’s why it’s defensible.

Engine #2: Attention yield (ads, partnerships, network monetisation)

This is the more contrarian piece — and therefore the more differentiating, if executed correctly.

Principle: If your product creates repeat engagement for a valuable audience, you can monetise the surface area — not by annoying users, but by enabling relevant, additive partner value.

The insight

Consumer platforms have understood attention economics for decades. Google and Meta built trillion-dollar businesses on attention monetisation.

But B2B software has traditionally ignored this revenue stream, viewing advertising as somehow beneath enterprise dignity. That’s changing. As subscription growth stalls, B2B companies are discovering that engaged user bases are valuable assets that can be monetised in multiple ways.

The cleanest framing: Software that owns attention can earn yield.

Yield can come from ads, partnerships, referrals, and transactions. The monetisation must be consent-driven, transparent, and value-adding.

In other words: don’t copy consumer adtech. Create utility-aligned monetisation.

Where the attention surface exists

Most B2B software never earns the right to monetise attention because users don’t choose to spend time there. But some categories do:

Surface	Why attention exists
Communication platforms	Daily, habitual engagement
Marketplaces and networks	Discovery and transaction intent
High-frequency dashboards	Operational necessity creates regular visits
Inbox-like environments	Daily attention is already there
Collaboration tools	Team workflows create repeated engagement

 How it works

Software that engages end-users creates an “attention surface.” That attention can be monetised through:

• Sponsored content from complementary vendors
• Partner revenue shares for referrals and transactions
• Contextual recommendations that add value
• Cooperative advertising networks where brands reach customers through each other’s engaged channels

The economics can be compelling. Instead of customers paying platform fees, the attention they generate subsidises their costs — while creating new revenue for the software provider.

The inbox example

Consider email — one of the few places where attention is both habitual and measurable.

Traditional model: brand pays email service provider per email sent, regardless of engagement.

Attention-based model: daily value-driven emails that customers actually want to open, with in-email transactions sponsored by complementary brands. The brand sending the email (publisher) earns revenue from the engaged attention. The brand reaching that audience (advertiser) pays for deterministic access — at a fraction of what they’d pay Google or Meta for probabilistic reach.

Both brands benefit. Attention is monetised. The network captures a share. This is the core of what I call NeoNet — a cooperative advertising layer built on authenticated engagement rather than probabilistic targeting. The economics favour everyone except the platforms currently extracting the reacquisition tax.

**

Why incumbents won’t copy easily

Even if these ideas sound straightforward, most incumbents struggle to execute because of structural barriers:

For performance fees:

• Sales compensation is built around ARR, not outcomes
• Finance teams dislike variable revenue until it becomes meaningful
• Product teams aren’t instrumented for causal measurement
• The shift requires confidence that the product actually delivers value — and willingness to stake revenue on that belief

For attention yield:

• “Ads” triggers cultural antibodies in B2B — even when it’s actually “partner yield”
• Engagement levels may not justify monetisation
• Network effects take time, and incumbents often have no incentive to start
• Different skill sets are required (media, partnerships, ad operations)

This is why the new engines require a distinct motion — not a side project for the existing team.

The moat: proprietary data and domain knowledge

Orlando Bravo is right that domain knowledge is the franchise. But he’s applying it defensively.

Applied offensively, domain knowledge becomes the foundation for new revenue models:

• Domain knowledge enables better outcome measurement. If you deeply understand a vertical, you know what “good” looks like. You can define baselines, measure uplift, and prove value in ways that generic AI models cannot.
• Proprietary data enables better targeting. If you have unique signals about customer behaviour, you can deliver more relevant recommendations and more valuable attention.
• Integration depth enables better execution. If you’re embedded in customer workflows, you can act on insights immediately.

The combination — proprietary data, domain knowledge, and new revenue models — creates defensible differentiation. Generic AI models can replicate features. They can’t replicate years of accumulated intelligence about specific verticals and customer behaviours.

The question for CEOs

A simple restatement:

The question isn’t “how do we defend subscriptions?”

It’s “what new profit pools can we add that scale with outcomes and attention?”

Now comes the hard part: execution without wrecking the core.

6

What B2B Software Companies Must Do

The two-track approach: protect the core, build the future

The biggest mistake companies make in moments like this is either:

• Freeze (“wait it out”), or
• Flail (a panicked “pivot” that scares customers and employees)

The right approach is a two-track operating system.

Why two tracks, not a pivot

The temptation in a crisis is dramatic action. Abandon subscriptions. Go all-in on outcomes. Transform the company.

This usually fails.

Subscriptions, for all their challenges, still generate predictable cash flow. That cash flow funds operations, pays salaries, and provides the runway to experiment. Abandoning it prematurely risks the entire company.

Existing customers chose you because of the current model. They have budgets allocated, contracts signed, workflows built. Suddenly changing the deal creates confusion and churn.

New models need time to prove. Performance fees require measurement infrastructure. Attention economics require engagement levels. Neither delivers revenue on day one. Betting everything on unproven models is reckless.

And the narrative matters. “We’re pivoting because our core business is failing” terrifies investors. “We’re building a second growth engine” inspires them.

The framing: Track 1 remains the compounding base. Track 2 is a deliberate build of new profit pools that escape seat compression. Add variable upside engines while there is still credibility, customers, and cash flow.

Track 1: Protect and modernise the core

Track 1 is the subscription engine. It funds everything.

Goals

• Reduce churn and contraction
• Improve product experience
• Use AI to lower cost-to-serve and increase value
• Tighten packaging so pricing feels earned
• Add AI features that justify (but don’t depend on expanding) current pricing

Metrics

Metric	What it measures
Gross retention	Customer base stability
Net revenue retention	Expansion vs. contraction
Logo churn	Customer loss rate
Gross margin	Operational efficiency
CAC payback	Sales efficiency
Product usage health	Real adoption, not vanity metrics

 Investment

The majority of current resources — 70-80%. Track 1’s job is to buy time and fund Track 2 experimentation.

Track 2: Build new revenue engines

Track 2 is not “innovation theatre.” It is a separate revenue motion with separate economics, skills, and success criteria.

Goals

• Prove outcome pricing in a tight set of pilots
• Build measurement credibility and repeatable playbooks
• Build (or plug into) an attention/yield network where the surface exists
• Convert pilots into a scalable revenue line with clear unit economics

Metrics

Metric	What it measures
Pilots launched	Experimentation velocity
Pilot → contract conversion	Model viability
Uplift distribution (median, p75, p90)	Outcome quality — not just best case
Time-to-proof	How fast you can show measurable value
Track 2 revenue as % of total	Revenue diversification progress

 Investment

A dedicated allocation — 20-30% of resources — with a dedicated team.

**

Organisational design: the anti-antibody move

The most common mistake is letting the Track 1 team run Track 2.

It never works.

Track 1 teams are optimised for subscription metrics. They’re compensated on ARR. They’ve spent careers perfecting the current model. Asking them to cannibalise it creates impossible conflicts.

Organisational antibodies are real. New initiatives that threaten existing revenue face resistance at every level. Priorities shift back to the core business. Experiments get deprioritised. Track 2 dies a quiet death.

The solution: structural separation

Dimension	Track 1	Track 2
Team	Existing organisation	Dedicated squad (startup-within)
Leader	Current leadership	Track 2 Lead with CEO visibility
Commercial motion	Standard contracts, quotas	Separate templates, legal, finance rules
Metrics	ARR, NRR, churn	Uplift, pilot conversion, Track 2 revenue
Compensation	Standard quotas	Tied to outcomes delivered / revenue realised
Culture	Optimise and defend	Experiment and prove

 The pilot programme: how to start without drama

Track 2 should start with pilots, not transformation.

Start with 3-5 pilots where:

• Baselines are measurable (you know current performance)
• The customer has executive sponsorship (champion who will advocate)
• The domain has clear value metrics (retention, conversion, cost savings)
• You can run controlled comparisons (treatment vs. control)

Pilot structure

Element	Specification
Duration	90 days + measurement window
Scope	10% of customer base in treatment, 10% in matched control
Methodology	Pre-agreed measurement protocol
Success criteria	Defined before starting

A good pilot is designed to answer one question: Can we repeatedly create measurable value, and can we contract for it?

Kill criteria: intellectual honesty

Avoid the sunk cost fallacy. Define failure upfront.

Track 2 should be reassessed if:

• Fewer than 30% of pilots show meaningful uplift
• Average uplift falls below the threshold that justifies the economics
• Track 1 materially suffers due to distraction
• 12 months pass without repeatable proof

Write these criteria down before starting. Revisit them at each gate. Be willing to kill what isn’t working.

The 90-day starting point

For companies ready to begin:

Weeks 1-2:

• Identify Track 2 leader (credibility, autonomy, CEO access)
• Define first pilot candidates from existing customers
• Draft measurement methodology

Weeks 3-4:

• Assemble core team (5-10 people, dedicated full-time)
• Sign 3-5 pilot agreements
• Establish baseline metrics

Weeks 5-8:

• Build/configure measurement infrastructure
• Deploy initial interventions
• Begin data collection

Weeks 9-12:

• First results emerging
• Iterate based on early learnings
• Expand to additional pilots if signals positive

Week 13+:

• Full measurement cycle
• Statistical significance achieved
• Decision: scale, iterate, or kill
• Board review

The investor narrative

Your outward story is not “we’re pivoting.” It’s: “We’re building a second growth engine.

Track 1 continues to deliver predictable subscription revenue and cash flow. We’re maintaining the business, adding AI capabilities, and extracting efficiency gains. This is the foundation.

Track 2 is capturing new profit pools through outcome-based pricing and attention economics. These revenue streams escape seat-based compression because they’re tied to value delivered, not headcount. They align our incentives with our customers’ success. And they position us for growth in the AI era.

We’re running both tracks in parallel. Track 1 funds the business. Track 2 builds the future. This isn’t a pivot — it’s diversification. We’re adding variable upside engines while we still have credibility, customers, and cash flow.”

What not to say:

• “We’re pivoting” (signals desperation)
• “Subscriptions are dead” (terrifies existing customers)
• “We’ll figure out the model later” (signals no plan)

The window is closing

The software companies that win the AI era won’t be the ones who cling longest to the old model.

They’ll be the ones who:

• Keep the subscription engine healthy, and
• Build new engines that monetise value delivered and attention earned

Enterprise software spending will exceed $1.4 trillion this year. Global IT spending will surpass $6 trillion. The market is growing. The opportunity is enormous.

The question is whether you’re positioned to capture it — or watching it flow to companies that figured this out faster.

The window is 18-24 months. After that, the market will reprice based on execution, not fear. Companies that have proven new revenue models will be rewarded. Companies that haven’t will face further compression.

The time to start is now.

**

Conclusion: The Revenue Imperative

B2B software’s crisis is real, structural, and accelerating.

The six threat vectors — seat slowdown, price increase backlash, AI budget shift, efficiency gaps, TAM traps, and product experience gaps — compound each other. The conventional responses — AI features, domain knowledge, cost cuts, consolidation — are defensive. They protect what exists. They don’t create what’s needed.

What’s needed is a new revenue playbook.

Performance fees that tie vendor revenue to customer outcomes. Attention economics that monetise engagement rather than headcount. New profit pools that escape seat-based compression.

The implementation requires a two-track approach. Track 1 maintains the subscription base, generates cash flow, and buys time. Track 2 builds new revenue streams, proves new models, and creates the future.

The question isn’t whether B2B software will change. It’s whether your company will lead the change or be changed by it.