Thinks 1774

Sam Altman on OpenAI’s hardware plans: “Two big revolutions in computer use have occurred: the mouse and keyboard, and the idea of the monitor displaying this sort of windowed system. That was a breakthrough, for sure. Then we had the touch devices, which adapted that, taking out the mouse and letting you use your finger, making it a very personal device. This was huge. Fundamentally, we have never had something as powerful as AI. Computers really can understand what we want, can think, which has let us reimagine what it could mean to use a computer. So we’re still exploring. It’ll take us quite some time. Don’t expect anything very soon. But over time, I expect we’ll make a small family of devices. They will look good, for sure, but that’s not the main thing. I hope that if we do a really great job, they will change what it means to use a computer, how you do your work, and how you play and live your life. But there’s a lot of work and a lot to explore between here and there.”

Richard Sutton: “First, the large language models are surprising. It’s surprising how effective artificial neural networks are at language tasks. That was a surprise, it wasn’t expected. Language seemed different. So that’s impressive. There’s a long-standing controversy in AI about simple basic principle methods, the general-purpose methods like search and learning, compared to human-enabled systems like symbolic methods. In the old days, it was interesting because things like search and learning were called weak methods because they’re just using general principles, they’re not using the power that comes from imbuing a system with human knowledge. Those were called strong. I think the weak methods have just totally won. That’s the biggest question from the old days of AI, what would happen. Learning and search have just won the day. There’s a sense in which that was not surprising to me because I was always hoping or rooting for the simple basic principles. Even with the large language models, it’s surprising how well it worked, but it was all good and gratifying. AlphaGo was surprising, how well that was able to work, AlphaZero in particular. But it’s all very gratifying because again, simple basic principles are winning the day.”

Andy Kessler: “Author Kyla Scanlon divides Generation Z into “safety seekers” and “digital gamblers.” Plausible, except careers that were once safe are now risky: graphic designers, marketers, some programmers, maybe even lawyers. And surfing the waves of progress to where the world is headed is less risky than you think. Progress comes via surprises, not rules, with inventions no one thought possible. The telescope opened the skies. The microscope illuminated the unseeable. Both surprises. So was Edison’s Kinetograph movie camera. Quantum theory was heretical, until it wasn’t. It enables entire industries, including semiconductors. Gene editing was hard until Crispr technology simplified it. Machine learning was researched for decades with little result, until back-propagation allowed voice and facial recognition. And it’s been less than three years since ChatGPT shocked the world with what it could do. None of these were invented by following the rules, but by coloring outside the lines…Take risks. Risk leads to reward. Ignore those who tell you to take “calculated risks.” It’s the magnitude of risk that provides the potential reward. And we need a new name for entrepreneurs. It’s too French. Maybe “risk agents” or “productivity creators” or, hmmm, “no rulers.””

Ashu Garg: “Your culture is defined by the people you hire and the behaviors you reinforce. The pace of AI startups is so fast that you need people who will truly do “whatever it takes.” To build this high-intensity culture, hire for an ownership mentality. Jonathan actively screens for what he calls “raw founder energy.” He hires “less for experience, more for exceptional ability…someone that’s really hungry, intense, hardworking.” In practice, this means Turing might pass on a polished veteran if a scrappier candidate shows more intensity and an eagerness to own outcomes. You want team members who care so deeply that they never consider any problem above or beneath them. That creates a culture where everyone pushes hard and actively works to make things better, rather than saying “that’s not my job.””

The Segment Martech Forgot: Why Rest Customers Hold the Key to Profitable Growth (Part 13)

Red Team

My colleague, Jaimit Doshi, created a CustomGPT to Red Team ideas. I used it on this essay.

🤔
– Thinker thoughts…
slip-slide of customers —
“daily games” won’t stop the leak;
data pipes still leak.

Short take: the essay (nice framing — Best → Rest → Test) spotlights a real and valuable problem: gradual disengagement is under-monitored and expensive to fix later. I agree the Rest deserves more attention. But the proposed NeoMarketing stack (Hooked Score + Brand Daily + BrandTwin + Rest Rescue + ActionAds + Progency + Alpha pricing) rests on a series of shaky technical, behavioural, regulatory and economic assumptions. If you build the full machine without testing those assumptions you’ll spend heavy CapEx and still watch customers quietly leave — only now you’ll have gamified the drain.

I’ll demolish the weakest assumptions, back the counter-claims with evidence, and finish with a pragmatic, prioritized experiment plan you can run in 90 days to prove whether this is a 10× opportunity or a sunk-cost fantasy. Where I cite the essay I’ll use your document; where I cite external facts I’ll use public sources.

Key claims from the essay (for context): the Rest segment is huge and under-served; daily 15–60s interactive touchpoints (Brand Daily) + zero-party data + BrandTwin give N=1 personalization at scale and monetisation via ActionAds/alpha pricing; Hooked Score will reliably predict Best→Rest transitions; Rest Rescue can be automated to restore 40–60% at a fraction of reacquisition cost.

Major red-team critiques (short + sharp)

1) Email/owned-channel plumbing is shakier than the essay assumes

The whole Brand Daily depends on high inbox placement + reliable interactive email capabilities (AMP or similar). But inbox providers and deliverability trends are volatile and fragmented:

  • Inbox placement and open metrics have been unstable across 2024–25; several providers tightened filters and Apple/Google privacy changes have disrupted open-rate signals. This undermines both reach and the “daily ritual” promise. (KL Communications)
  • AMP-for-email adoption remains niche, inconsistent across providers, and has a history of uneven support (Microsoft experimented then pulled support; developer complexity and deliverability concerns remain). Betting a large product on AMP’s universal availability is risky. (Buttondown)

Consequence: The Brand Daily may not reach the audience reliably; interactivity may only work for a subset of users or clients. That breaks the core “15–60s habit” assumption.

2) Zero-party data is helpful but overhyped and biased

You argue BrandTwin’s zero-party data will deliver accurate N=1 profiles. Reality: self-reported preferences are valuable but noisy and systematically biased (social desirability, low introspective accuracy, survey fatigue). Academic and industry work warns zero-party is not a magic bullet — it complements, not replaces, behavioral (first-party) data. (ResearchGate)

Consequence: You’ll get enriched profiles — but they won’t always predict purchase behaviour. BrandTwin must fuse zero-party with behavioral signals and validate predictions continuously; otherwise personalization will misfire and annoy customers.

3) Gamification / rewards (Mu) → fatigue & marginal utility

Gamification can boost early engagement, but evidence shows novelty wears off and over-gamification produces fatigue and even stress for some user groups. Rewards change incentives — you risk building an “open-for-points” audience rather than a loyal buying audience. (ScienceDirect)

Consequence: Mu could inflate engagement metrics but not meaningfully increase LTV — or worse, create a cohort that only engages for freebies and then lapses.

4) Monetising sends (ActionAds, brand-to-brand ad networks) creates privacy, brand-fit, and trust risks

The essay’s “ActionAds / NeoN” idea — monetising every send via third-party sponsorship — sounds attractive but raises red flags:

  • Identity-sharing across brands to target audiences is constrained by GDPR/CCPA-style rules and recent reductions in ID-syncing after GDPR. Legal complexity and customer trust fallout are real. (rub.de)
  • Sponsorships risk ruining the “non-promotional” promise of the Brand Daily: once users see marketing from other brands in a supposed “relationship” touch, mental salience may shift to the sponsor, not you. That’s a cannibalisation risk.

Consequence: Monetisation may undercut the relationship-building value proposition and introduce compliance overhead or reputational risk.

5) Hooked Score is elegant — but metric-design and causality are hard

Hooked Score’s design (points for opens/clicks, streak multipliers, decay) is sensible, but:

  • It risks metric gaming (optimize for actions that increase score without changing purchase propensity).
  • Engagement signals (open/click) are noisy proxies for intent, especially after privacy changes; places where the essay substitutes opens for true interest are fragile. (Validity)
  • The assumed rescue success rates (40–60%) are optimistic without controlled experiments.

Consequence: If Hooked Score triggers interventions based on noisy signals, you will waste Mu and human attention on low-lift cases and possibly annoy customers.

6) Operational & organizational friction is understated

You call out silos — good — but your solution requires cross-org alignment (product, customer service, privacy/legal, data engineering, paid media, finance). The Progency/Alpha pricing pitch attempts to sidestep this, but outsourcing can’t remove strategic dependencies: data access, identity resolution, billing, inventory, and CX ownership. Outsourcing without tight SLAs and governance often produces finger-pointing and poor long-term outcomes.

Consequence: Implementation failure is likely unless the vendor model is iron-clad and the organisation dedicates accountable owners.

7) Regulatory and future-proofing gaps

  • Data localization and new privacy bills (India, Brazil, EU, US states) make cross-border identity-sharing and cooperative ad networks complex.
  • Zero-party data plus active monetisation must be carefully consented — or you risk fines and brand damage.

Evidence that supports the essay’s premise (so I’m not tossing the baby out)

  1. Retention is usually cheaper than acquisition — long-standing industry rule-of-thumb: retention can be multiple times cheaper than acquisition. (Use as directional, not exact multiplier.) (Invesp)
  2. Attention scarcity + marketing fatigue is real; customers increasingly block/ignore irrelevant messaging. That strengthens the argument to focus on salience and habit. (Optimove)

So the problem framing is solid; the solution suite needs rigorous validation and risk management.

Concrete fixes & a 90-day red-team experiment plan (do these BEFORE you build the full stack)

Goal: validate 3 critical assumptions quickly and cheaply:
A. Do lightweight daily touchpoints raise predictive purchase propensity vs. control?
B. Can zero-party signals + behavioral data improve short-term LTV predictions beyond current models?
C. Is monetisation via sponsorships net-positive for relationship and margin?

Design three parallel experiments.

Experiment A — “Mini Brand Daily” A/B test (8 weeks)

  • Population: 50k customers split: Control (current CRM) vs Test (mini Brand Daily).
  • Create a single channel lightweight daily micro-touch (no AMP required): use simple responsive HTML + progressive micro-interactions (poll + single CTA) that degrade gracefully for non-AMP clients. Limit to customers with past 6-12 mo LTV in Rest band.
  • Metrics (primary): 90-day purchase rate lift (indicator), incremental revenue per user (IRPU). Secondary: unsubscribe rate, spam complaints, net promoter delta.
  • Duration: 8 weeks active + 8 weeks measurement.
  • Acceptance criteria to scale: statistically significant +3–5% increase in 90-day purchase probability and no >0.2% increase in spam complaints.

Why this matters: proves the behavioral hypothesis without building BrandTwin or Hooked Score. Use simple A/B to avoid confounding.

Experiment B — “Hooked-lite” predictive test (8–12 weeks)

  • Build a pared-down HookedScore that uses robust signals only (site visits, purchases, time-on-site, CTRs) — avoid opens as a primary signal due to privacy noise.
  • Train a model to predict 90-day spend/classify “at risk” vs baseline RFM model. Evaluate uplift in early detection (precision/recall) and cost per saved customer assuming simple re-engagement (one email + small Mu).
  • Metrics: ROC-AUC improvement vs baseline RFM; uplift in predicted LTV; real-world rescue conversion rate for the predicted at-risk cohort.
  • Acceptance: model must materially (e.g., +10–20% AUC or +15% better recall at same precision) beat RFM and yield rescues with positive ROI.

Why this matters: proves Hooked Score’s incremental predictive power before you deploy automated Rest Rescue.

Experiment C — Monetisation pilot & brand-fit test (8–12 weeks)

  • Offer a small cohort of Brand Daily slots to highly non-competing, vetted sponsors (one per cohort). Use clear opt-in and separate consent for sponsored content.
  • Track revenue per send, direct response to sponsor, and any change in primary brand conversion or trust signals (brand lift surveys).
  • Metrics: net revenue per send, net effect on primary-brand LTV, opt-out delta, compliance/legal flags.
  • Acceptance: monetisation should not reduce primary-brand conversion or increase opt-outs materially; sponsors must provide >X revenue per 1k sends to offset marginal cost.

Why this matters: validates the ActionAds/economic model without full NeoN rollout.

Tactical changes to the proposed architecture (if you continue)

  1. Progressive rollout — build the Brand Daily to degrade gracefully; do not require AMP. Use in-email links to compact micro-experiences on fast landing pages optimized for one click. This reduces deliverability/compatibility risk. (Essay assumes AMP everywhere; don’t.) (Buttondown)
  2. Fuse zero-party with behavioural data — require BrandTwin to treat zero-party fields as signals not ground truth. Validate with observed purchase behaviour and adapt weights. (Research warns self-report bias.) (ResearchGate)
  3. Design Mu to be loss-averse and tied to purchase — avoid pure extrinsic rewards that create “points-only” behavior. Use Mu as a pairing mechanism: small MUs for engagement, but larger MU multipliers when engagement leads to purchase or referral.
  4. Make Hooked Score auditable & anti-gaming — require attribution that ties score rise/fall to downstream revenue lifts. Track false positives (alerts where rescue failed) to tune decay/weights.
  5. Privacy-by-design & legal gating — any cross-brand monetisation must be explicit opt-in, with clear data flows and contractual guarantees about data use and localization. (rub.de)
  6. Org ownership — create a single cross-functional “Retention & Relationship” owner with a clear P&L and KPI that spans CRM, product, support, and paid channels. Without that, the silo problem remains.

Metric dashboard you should require before scaling

  • Net Revenue per Rest User (NRRU) (30/90 day windows)
  • Rescue lift (%) and cost per rescued customer
  • Unsubscribe & spam complaint delta for Brand Daily sends
  • BrandDaily stickiness: active daily interaction rate and streak decay
  • Monetisation revenue per 1k sends & sponsor fit index
  • Model precision/recall vs RFM baseline (Hooked Score)
  • Privacy/consent friction metrics (consent drop-off rate, legal incidents)

**

Final verdict (short)

The essay’s diagnosis is strong — Rest is a real opportunity. The prescription is promising but far too “big-bet” without staged validation. The biggest risks are channel fragility (email deliverability & AMP), over-reliance on zero-party data, gamification fatigue, and compliance/monetisation tradeoffs. Do the 90-day experiments above. If they hit your acceptance criteria, scale carefully. If not, iterate: shorten the Brand Daily, change reward mechanics, broaden signal sources.

Parting statesman (funny & wise)

You’re asking marketing to become a lighthouse keeper: keep the light burning so ships (customers) don’t run aground. Don’t spend the kingdom’s treasure building a giant, ornate lighthouse with gold roofs (AMP chandeliers, Mu fountains, NeoN marketplaces) before you’ve checked the coastline for fog. Test whether a humble lamp on a sturdy post — a simple daily note, a clean predictive flag, a respectful ask for a preference — actually keeps ships near. If it does, then build the cathedral. If it doesn’t, at least you saved the kingdom from a very expensive, very shiny, and rather pointless cathedral.

Thinks 1773

NYTimes: “Competition in China is often far more cutthroat than in the United States. America has a handful of carmakers; China has more than 100 electric vehicle makers struggling for market share. China has so many solar panel makers that they produce 50 percent more than global demand. About 100 Chinese lithium battery producers churn out 25 percent more batteries than anyone wants to buy. This forces Chinese manufacturers to innovate, but it also leads to price wars, losses and bad debt — and that’s becoming a problem. China is heading toward deflation, the often catastrophic downward spiral of prices that sank Japan in the 1990s. Its leaders are blaming a culprit they call “involution” (“neijuan” in Mandarin), a term that has come to mean reckless domestic competition. They want to rein it in by browbeating companies into keeping prices steady and instructing local governments to scale back subsidies.”

Study Finds: “Defiance is not about being oppositional for its own sake. It’s about choosing to act in line with your values when there is pressure to do otherwise. That pressure can come from anywhere: a boss urging you to fudge the numbers, a friend nudging you toward something you don’t believe in, a culture telling you to stay in your place. Defiance in those moments might be as small as saying “no,” asking for clarification or simply pausing instead of rushing along with the group. Other times, it means speaking up, challenging authority or maybe walking away. Seen this way, defiance isn’t a fixed trait that some people are born with and others lack. It’s a practice: a skill you can strengthen over time. Some days you might comply, other days you might resist. What matters is that you have the awareness and the tools to make the choice consciously, rather than letting fear or habit decide for you.”

Mint: “I’m no longer sure if the train remains as much of a social melting pot of chatter and exchange, when our closest companions are handheld devices and watching reels without earphones is our favourite pastime. My last few train journeys—mostly short ones by chair cars—were memorable for the loud phone conversations that sometimes went for the entire duration of travel or wailing children being placated by electronic devices. Noise cancelling headphones are now a new normal, which not only drown out the ambient chaos but, sadly, also the clickety-clack of the wheels. [Monisha] Rajesh, too, admits that the quality of connections on trains have changed in the last 15 years she has been writing about her journeys. “I avoid using my phone on trains, except as a recording device or to take an occasional photograph. Ironically, now, even those who are taking trains because they want to experience ‘slow travel’ seem to be completely sucked in by their phones,” she says.” More: “Two hundred years after the first train plied, train travel endures despite being written off every now and then. The sway of carriages, the changing scenery through the windows and chance encounters make it the most enchanting form of travel.”

WSJ: “The key is enabling AI to learn from their environments and faithfully represent an abstract version of them in their “heads,” the way humans and animals do. To do it, developers need to train AIs by using simulations of the world. Think of it like learning to drive by playing “Gran Turismo” or learning to fly from “Microsoft Flight Simulator.” These world models include all the things required to plan, take actions and make predictions about the future, including physics and time. The world-model approach—which somewhat confusingly refers to both the simulated training environment and the abstract representation—is already having potentially huge effects on the real world. Drone warfare, new kinds of robots and safer-than-human self-driving vehicles all benefit from it, says Moritz Baier-Lentz, a partner and investor at Lightspeed, a venture-capital firm. There’s an almost unanimous belief among AI pioneers that world models are crucial to creating next-generation AI.”

The Segment Martech Forgot: Why Rest Customers Hold the Key to Profitable Growth (Part 12)

Story: ChatGPT

The late afternoon light slanted through the glass panels of the conference room, reflecting off the giant dashboard on the wall. Rows of numbers pulsed in real time—campaign clicks, ROAS ratios, acquisition spends. Maya, the CMO of a leading D2C beauty brand, leaned back in her chair, rubbing her temples.

“Seventy per cent of our budget goes to performance marketing,” she muttered. “And yet, every quarter, we start from zero again.”

Vijay smiled. He’d heard this line before—different company, same problem. He slid his tablet across the table. “That’s because you’re chasing ghosts,” he said.

Maya frowned. “Ghosts?”

“Your own customers,” he replied. “People you already had. You’re paying to reacquire them after they quietly drift away. That’s the leak every brand has—but no one tracks it.”

He tapped his screen, and a visual appeared: three circles labelled Best, Rest, and Test. “You’ve got your Best customers—the loyalists who buy again and again. And your Test—those who’ve vanished. But what about the middle?”

“The Rest?” Maya asked.

“Exactly. They’re the ones who used to love you but are now slipping away. You don’t see them because your systems measure conversion, not disengagement. By the time you notice, they’re gone—and Google’s charging you to get them back.”

Maya crossed her arms. “We already have reactivation campaigns.”

“Campaigns?” Vijay smiled. “That’s the problem. Campaigns talk at customers. The Rest need relationships. You can’t fix fading engagement with a discount. You fix it with presence.”

He swiped again. “Here’s how NeoMarketing does it. Step one: we track the Best→Rest transition using something called a Hooked Score. It measures micro-engagements—opens, clicks, streaks, small interactions—and alerts you when loyalty starts to fade. You don’t wait for churn; you see it coming.”

She leaned forward. “And then?”

“Step two: you engage them with The Brand Daily. A 15–60 second daily email ritual. No hard selling—just micro-games, tips, quizzes, useful content. A reason to open, not an offer to ignore. It keeps attention alive and the brand front-of-mind. And we make it pay for itself with ActionAds or Alpha-based pricing. Even Rest engagement becomes profitable.”

Maya smiled sceptically. “You think people will open an email every day?”

“They open Instagram every day,” Vijay countered. “Attention isn’t the problem; relevance is. The Brand Daily feels like entertainment, not marketing. And as customers engage, we start building their BrandTwin—an AI profile that learns from their responses. Every day, it gets smarter, so the content feels more personal. Over time, it moves from general to one-to-one.”

“So it’s like having a personalised marketing assistant for each customer?”

“Exactly. Except it works automatically. The twin learns, nurtures, and nudges—before disengagement turns into defection. And when engagement drops sharply, Rest Rescue kicks in—tiered interventions that range from gentle nudges to personalised outreach. No one falls through the cracks.”

Maya studied the diagram. “So Hooked Score shows me who’s slipping. The Brand Daily keeps them hooked. BrandTwin makes it personal. And Rest Rescue saves the ones about to leave.”

Vijay nodded. “You got it. You stop the leaks before they become floods. You save the Rest, and you save your profits. No more paying Google to reacquire the customers you already owned.”

Maya was quiet for a moment, then said softly, “All this time, we’ve been measuring transactions, not transitions.”

“Exactly,” said Vijay. “NeoMarketing flips that lens. It’s not about buying more attention. It’s about keeping the attention you already earned.”

She smiled, eyes narrowing in thought. “Alright, Vijay. Let’s bring my Rest back to life.”

Thinks 1772

Ben Thompson: “Every item in every YouTube video is well on its way to being a monetizable surface. Yes, that may sound dystopian when I put it so baldly, but if you think about it you can see the benefits; I’ve been watching a lot of home improvement videos lately, and it sure would be useful to be able to not just identify but helpfully have a link to buy a lot of the equipment I see, much of which is basically in the background because it’s not the point of the video. It won’t be long until YouTube has that inventory, which it could surface with an affiliate fee link, or make biddable for companies who want to reach primed customers.”

Bain Technology Report 2025. One of the questions discussed is Agentic AI vs SaaS. “Generative and agentic AI are disrupting software as a service (SaaS) by automating tasks and replicating workflows. SaaS leaders can manage the risks by identifying where AI can enhance their offerings and where it might replace them. To stay ahead, they must own the data, lead on standards, and price for outcomes, not log-ons, in an AI-first world. With the right playbook that includes deep AI integration, strong data moats, and leadership on standards, incumbents can shape, not just survive, the next wave of SaaS.”

WSJ: “Global companies looking to build factories in India appreciate the country’s vast workforce and its potential as a consumer market, but they are often deterred by bureaucracy. The obstacles aren’t just about setting up a business—where officials have recently tried to smooth the path—but also about closing one down if it doesn’t pan out. Regulations make layoffs difficult, politicians are often loath to see investors leave and courts can issue wildly contrary rulings even when the facts are similar. “Exit barriers are a key reason for India’s underdeveloped manufacturing sector,” said Shoumitro Chatterjee, co-author of a new paper on Indian manufacturing and an assistant professor of international economics at Johns Hopkins University. It “acts as an entry cost and deters potential companies.””

Bloomberg: “AI is set to change the way we pay for almost everything. From airline tickets to online purchases, prices could soon be tailored for you—lower or higher than the average. With personalized pricing driven by know-all algorithms, how badly you need something would be hard to conceal.”

The Segment Martech Forgot: Why Rest Customers Hold the Key to Profitable Growth (Part 11)

Story: Claude

Maya stared at her laptop screen, the quarterly marketing review spreadsheet mocking her with red numbers. Revenue up 12%, marketing spend up 43%. The CFO’s email sat unopened in her inbox—subject line: “We need to talk about ROI.”

Her assistant knocked. “Vijay’s here for your 3 o’clock.”

Maya had almost cancelled. Another martech vendor promising magical solutions she’d heard a hundred times before. But Vijay’s email intrigued her: “I’m not here to sell you software. I’m here to show you where your profits are bleeding—and how to stop it.”

Vijay walked in, declined coffee, and opened not a pitch deck but Maya’s own customer database on his laptop.

“Maya, quick question: what percentage of your customers are in each segment—Best, Rest, Test?”

She pulled up her CRM dashboard. “About 22% Best, 38% active but not top-tier, and 40% dormant or churned.”

“That middle 38%—you call them ‘active but not top-tier.’ We call them Rest. Can you show me your campaigns targeting them specifically?”

Maya scrolled through her campaign calendar. “Well, they’re included in our promotional emails, seasonal campaigns—”

“Same campaigns as your Best customers?”

“Mostly, yes. Slightly different offers based on purchase history.”

Vijay leaned forward. “Here’s what your data actually shows.” He pulled up a visualisation she’d never seen. “Three months ago, you had 28% Best customers. Today, 22%. Where did they go?”

Maya’s eyes widened. “They… moved to Rest?”

“Exactly. Six percentage points—about 24,000 customers—slid from Best to Rest in one quarter. Your average Best customer generates ₹45,000 annual revenue. These 24,000 just cost you ₹108 crores in future revenue. And in another 90 days, most of them will hit Test. Then you’ll spend 5-10x trying to win them back through adtech.”

The number hit Maya like cold water. “Why didn’t my martech platform alert me to this?”

“Because it measures conversions, not transitions. It tracks yesterday’s success, not tomorrow’s bleeding. Your Best customers don’t churn overnight—they drift. First they open fewer emails. Then they skip purchases. Then they’re gone. And because you’re measuring segment averages, not individual trajectories, the transition is invisible until it’s too late.”

Maya sat back. “So what’s your solution? More aggressive retention campaigns? Deeper discounts?”

“Worse than wrong—that accelerates the slide. Rest customers aren’t leaving because your offers aren’t good enough. They’re leaving because you forgot they existed. They have category loyalty, not brand loyalty. When they need grocery delivery next month, they’ll choose whoever’s most present in their mind. Right now, that’s not you.”

Vijay opened a demo. “The Brand Daily. Sixty seconds. Every day. Not selling—connecting. A cooking quiz earns them Mu points. A recipe tip. A poll about meal preferences. Three personalised product suggestions. Maybe an ad from a complementary brand that generates revenue even when they don’t buy from you. Frictionless, valuable, habit-forming.”

“You want me to email customers daily? They’ll unsubscribe.”

“Only if you’re boring. If you’re useful, entertaining, rewarding? They’ll expect it. Demand it. Your Hooked Score—our engagement metric—will tell you exactly when someone’s fading. Drop 20%? Gentle nudge. Drop 40%? Change content format. Drop 60%? Personal intervention. We catch them in Rest, not Test.”

Maya was sceptical. “This sounds expensive. My CFO is already questioning my budgets.”

“That’s why we use Alpha pricing. We only get paid when you get measurable results above baseline. If we don’t move Rest customers back to Best or prevent Best→Rest slides, we make nothing. We’re not a vendor—we’re a partner betting on your success.”

“And the daily content creation? My team doesn’t have bandwidth.”

“That’s Progency. We’re your outsourced Rest department. Platform, experts, AI agents, continuous improvement—all focused on the segment you don’t have time to manage. Your team focuses on Best customers and acquisition. We handle Rest. Different muscles, different skills.”

Maya looked at her spreadsheet again, seeing it differently now. Those 24,000 customers who slid from Best to Rest weren’t lost yet. They were saveable.

“Show me the ROI model.”

Vijay smiled. “Conservative assumptions: we save 40% of Best→Rest transitions. Each saved customer preserves ₹45,000 annual value. 24,000 × 40% × ₹45,000 = ₹43 crores recovered annually. Cost: ₹8 crores in Progency fees. That’s 5.4x ROI—and we’re eliminating the reacquisition waste when they hit Test.”

He pulled up one more screen. “But here’s the real magic. BrandTwin—their AI digital twin—learns from every Brand Daily interaction. Zero-party data. Preferences, context, timing, content affinity. After 90 days, we’re not personalising to segments of thousands. We’re personalising to segments of one. Your Rest customers feel seen, known, remembered. When they’re ready to buy again, there’s no question who they choose.”

Maya did the maths in her head. Forty-three crores recovered. Maybe another twenty crores in prevented reacquisition waste. Sixty-three crores total impact for eight crores investment. Her Rule of 40 problem suddenly looked solvable.

“When can you start?”

“First, we need to set up Hooked Score tracking—about two weeks of implementation. Then we launch Brand Daily with your top 50,000 Rest customers—controlled test vs. baseline. In 30 days, you’ll see the data. In 60 days, you’ll see the revenue. In 90 days, you’ll wonder why you ever let martech convince you that Rest customers didn’t matter.”

Maya opened the CFO’s email and hit reply: “Let’s move our next meeting. I have a solution to show you.”

Vijay packed his laptop. “One more thing, Maya. This isn’t just about saving Rest customers. It’s about transforming how you think about marketing. From cost centre hoping for results to profit engine delivering guarantees. From broadcasting to relationships. From segments to individuals. From fighting fires to preventing them.”

He paused at the door. “Your competitors are still playing the old game—acquire, campaign, lose, reacquire, repeat. You’re about to play a different game entirely. Welcome to NeoMarketing.”

After he left, Maya pulled up her customer list and filtered for Hooked Scores below 50. Thousands of names appeared—customers she’d been ignoring, customers drifting away, customers she could still save.

The segment martech forgot was finally getting the attention it deserved.

Thinks 1771

Mint: “Investors have historically underestimated the total addressable market that Indian startups can disrupt, often failing to envision the scale companies can achieve in India, Rajan Anandan, managing director at Peak XV Partners, told Mint in an interview. “I think all of us in India, including us, underestimate how big something can get here. A decade ago, if a consumer brand scaled to ₹100 crore, everybody said ‘wow’. Now, ₹100 crore is nothing — people ask if it can get to ₹1,000 crore or even ₹8,000 crore. A decade from now, we’ll be talking about ₹10,000-crore companies,” he said. Anandan said every decade in India brings a 10-fold jump, which makes it difficult for any investor to envision outcomes. “If you ask me what we underestimated, it’s the scale of India. An India at $7-8 trillion GDP will create companies far bigger than those that can be created at $4 trillion GDP. That’s what to watch out for,” he said.”

McKinsey: “Achieving business value with agentic AI requires changing workflows. Often, however, organizations focus too much on the agent or the agentic tool. This inevitably leads to great-looking agents that don’t actually end up improving the overall workflow, resulting in underwhelming value. Agentic AI efforts that focus on fundamentally reimagining entire workflows—that is, the steps that involve people, processes, and technology—are more likely to deliver a positive outcome. Understanding how agents can help with each of these steps is the path to value. People will still be central to getting the work done, but now with different agents, tools, and automations to support them. An important starting point in redesigning workflows is mapping processes and identifying key user pain points.”

SaaStr: “The tough, honest start-up truth: 1% of your team is epic, 5% of your team is great, 15% of your team gets sh*t done. 50% of your team does OK-to-solid work, 30% of your team adds little value at all. This has honestly always been true.  It’s just in 2021, there was so much organic demand for everything in business software, it masked things.  And the 50% and midpack could, e.g., still be decent closers. Not today. AI is coming for the last two groups. Faster than most want to be honest about.”

Dean Ball: “A core problem about the thing we today call “artificial intelligence policy” is that different people mean radically different things by the phrase “artificial intelligence.” I mean the term “radically” here in its literal sense: there are some who believe that AI systems capable of end-to-end autonomy of all human cognitive labor are coming soon, and there are others who assert, either explicitly or implicitly, that this is impossible, or who otherwise refuse to take the idea of this technology being built seriously. Both groups are basically bullish: the latter group believes in “really good LLMs,” and the former group believes in “AGI,” “transformative AI,” “powerful AI,” and various other terms used to refer to future AI systems. This creates a policy planning problem.”

WSJ asks if Gen Z is unemployable: “Hiring managers prize achievement, learning and work. Today’s youth value pleasure and individuality.”

The Segment Martech Forgot: Why Rest Customers Hold the Key to Profitable Growth (Part 10)

Summary

For decades, marketers have managed customers through a binary lens: those who are actively engaging and buying (the Best) and those who have become dormant or churned (the Test). The Best receive campaigns and loyalty schemes, whilst the Test are targeted with offers, discounts, and expensive reacquisition via adtech. This binary model has dominated CRM playbooks, but it misses the truth: churn is not sudden, it is a gradual slide. Customers don’t leap from Best to Test overnight—they transition.

That transition state creates the Rest segment: customers drifting away, neither loyal nor lost, representing marketing’s most profitable missed opportunity. Rest customers account for 30% of revenue from 40% of the base—three times more valuable per customer than Test—yet they receive the least attention. Marketers pour resources into wooing the Best and rescuing the Test, whilst the Rest quietly leak away.

Why have marketers overlooked the Rest? Because martech platforms track conversions, not transitions. Dashboards celebrate what works but ignore what’s quietly breaking. Adtech made reacquisition deceptively easy, luring marketers away from long-term relationship building. CRM lacked tools for relationship content beyond campaigns and discounts. Organisational silos meant no team owned the transition, whilst KPIs rewarded short-term revenue, not churn prevention. Segmentation concealed micro-patterns by averaging millions together. Until recently, technology and cost constraints made it difficult to track engagement signals or deliver lightweight daily content at scale.

But the world has changed. The acquisition pool is drying up; “new” customers are mostly reacquired defectors or expensive competitive switchers. Profitability pressures demand Rule of 40 performance—more value from existing customers and less waste on reacquisition. Customers are blocking bland, repetitive campaigns in an attention-starved world. Simultaneously, breakthroughs in AMP for interactive email, generative AI for content creation, and AI Agents for monitoring have made Rest management technically possible. Business model innovations—ActionAds monetisation, Progency outsourcing, and Alpha outcome-based pricing—make it economically viable. The result is a perfect storm: ignoring Rest is no longer a blind spot, it is reckless economics.

NeoMarketing offers a systematic solution through four integrated breakthroughs:

  • Hooked Score tracks the Best→Rest transition through a dynamic engagement index based on opens, clicks, streaks, and interactions across multiple time horizons (30/90/180 days). It acts as an early-warning radar that traditional martech lacks, detecting decline before it becomes churn.
  • The Brand Daily engages Rest customers with a 15-60 second daily ritual delivered through owned channels. Powered by Microns (brain games), Mu rewards (gamification), zero-party data collectors, and ActionAds, it preserves attention, builds mental salience, and monetises engagement even when purchases don’t happen. It’s not promotion—it’s presence.
  • BrandTwin builds an AI-driven digital twin for every customer that progressively thickens data files through zero-party data collection. In its Passive phase, BrandTwin experiments and learns preferences. In its Active phase, it delivers nudges and next-best actions, progressing from N=few to N=1 personalisation. The flywheel accelerates: engagement generates data, data improves personalisation, personalisation drives engagement.
  • Rest Rescue provides automated intervention workflows triggered by Hooked Score declines. Tiered protocols escalate from gentle nudges (15-20% score drop) to emergency intervention (>60% drop), combining AI automation with human escalation for high-value customers. It systematically saves declining relationships before they reach expensive Test status.

The shift in philosophy is simple yet profound: Best customers need retention campaigns; Rest customers need relationship maintenance. This means moving from transaction pressure to presence, from chasing conversions to creating mental salience, from recognition to remembering that most customers have category loyalty, not brand loyalty.

Together, these four solutions detect, engage, personalise, and rescue Rest customers before they become expensive reacquisition problems. They transform marketing from reactive cost centre to proactive profit engine. In the invisible transitions of the Rest lies marketing’s biggest hidden opportunity. Save the Rest, and you save the profits. Ignore them, and you bleed endlessly into AdWaste. NeoMarketing transforms this overlooked middle from silent leakage into the foundation of profitable, sustainable growth.

Thinks 1770

Economist: “The secret of economic success is innovation and the secret of innovation is the brilliance, creativity and drive of the most talented few. But even as governments throw money at schemes to boost their economies, including chipmaking factories and rare-earth mines, brainpower is going unloved. And the waste is getting worse.”

Tomasz Tunguz: “We build teams in pyramids today. One leader, several managers, many individual contributors. In the AI world, what team configuration makes the most sense? Here are some alternatives : First, the short pyramid. Managers become agent managers. The work performed by individual contributors of yore becomes the workloads of agents. Everyone moves up a level of abstraction in work. Second, the rocket ship! One director, seven managers, 21 employees. Everyone in the organization is managing agents, but these agents reflect their seniority. The director manages an AI chief-of-staff, the managers are player-coaches, both executing goals themselves & training/coaching others on how to manipulate AI successfully, which cuts the span of control by half. This configuration reduces headcount (1:7:49 -> 1:7:14) by 53%.”

WSJ: “The artificial-intelligence boom has ushered in one of the costliest building sprees in world history. Over the past three years, leading tech firms have committed more toward AI data centers like the one in Ellendale, plus chips and energy, than it cost to build the interstate highway system over four decades, when adjusted for inflation. AI proponents liken the effort to the Industrial Revolution. A big problem: No one is sure how they will get their investment back—or when.”

NYTimes: “China has embarked on a campaign to use more robots in its factories, transforming its manufacturing industries and becoming the dominant maker…There were more than two million robots working in Chinese factories last year, according to a report released Thursday by the International Federation of Robotics, a nonprofit trade group for makers of industrial robots. Factories in China installed nearly 300,000 new robots last year, more than the rest of the world combined, the report found. American factories installed 34,000. While Chinese factories have been using more robots, they have also gotten better at making them. The government has used public capital and policy directives to spur Chinese companies to become leaders in robotics and other advanced technologies like semiconductors and artificial intelligence.”

The Segment Martech Forgot: Why Rest Customers Hold the Key to Profitable Growth (Part 9)

Rest Rescue: Automated Intervention Workflows

The Problem We’re Solving:
Detecting the Best→Rest transition through Hooked Score is valuable, but detection without intervention is merely interesting. What specific actions should trigger when a customer’s score drops? How should intervention intensity scale with decline severity? Who’s responsible for execution? Without systematic protocols, insights become shelf-ware.

The Rest Rescue Solution:
Rest Rescue is a tiered, automated intervention system that activates when Hooked Score signals Best→Rest transition. It combines AI-powered detection, progressive intervention protocols, and human escalation pathways to systematically save declining relationships before they reach expensive Test status.

The Tiered Intervention Framework:
Rest Rescue operates on escalating intervention levels triggered by Hooked Score thresholds:

Level 1: Gentle Nudge (Score drops 15-20%) Trigger: 30-day Hooked Score declining but still above Rest threshold
Intervention:

  • Add Atomic Reward (Mu) bonuses to incentivise re-engagement
  • Deploy “We miss you” content featuring best-performing historical topics
  • Activate BrandTwin’s “win-back” content library
  • No human involvement—fully automated

Level 2: Active Re-engagement (Score drops 20-40%) Trigger: Customer now in Rest territory, declining trajectory
Intervention:

  • Switch to high-value content (educational deep-dives, exclusive access)
  • Offer “streak revival” Mu bonus to rebuild habits
  • Deploy preference update request: “Help us serve you better”
  • Activate multi-channel approach (email + SMS/WhatsApp)
  • BrandTwin experiments with different content formats
  • Alert account manager for high-LTV customers

Level 3: Intensive Intervention (Score drops 40-60%) Trigger: Deep Rest, approaching Test threshold
Intervention:

  • Personal outreach: Founder/CEO email (“We’d hate to lose you”)
  • VIP offer: Exclusive discount, early access, or surprise gift
  • Direct feedback request: “What did we do wrong?”
  • Human touch: Call from customer success for high-value customers
  • Channel preference reset: “How do you prefer to hear from us?”
  • BrandTwin creates custom recovery journey based on historical engagement peaks

Level 4: Emergency Protocol (Score drops >60% or enters Test) Trigger: Customer has crossed into Test territory
Intervention:

  • “Last chance” re-engagement series
  • Maximum-value offer (significant discount, bundle, or exclusive)
  • Multi-channel blitz (email, SMS, push, retargeting via NeoN)
  • Executive intervention for strategic accounts
  • Exit survey with incentive: “Tell us why you’re leaving—₹500 credit for honest feedback”
  • Data preservation: Archive BrandTwin for potential future reactivation

Channel Shifting Logic:
Rest Rescue doesn’t just vary content—it shifts channels based on engagement patterns:

Primary Channel Declining (email engagement drops):

  • Switch to SMS/WhatsApp for immediate touchpoints
  • Activate push notifications (app)
  • Test postal mail for high-LTV (yes, physical mail still works)
  • Deploy retargeting via NeoN (owned ad network)

Multi-Channel Orchestration:

  • Coordinate message timing across channels
  • Ensure consistent narrative (not repetitive spam)
  • Respect frequency caps per channel
  • Track cross-channel engagement for Hooked Score

The Progressive Commitment Ladder:
Rest Rescue interventions follow a psychology-backed progression:

  • Low-friction asks first: “Just open this email” (Mu reward)
  • Small commitments: “Tell us one preference” (Mu + better personalisation)
  • Medium engagement: “Play this 30-second game” (Mu + entertainment value)
  • Higher investment: “Complete this survey” (Mu + tangible benefit)
  • Transaction trigger: “Exclusive access just for you” (special offer)

Each step rebuilds engagement muscles and re-establishes habits before asking for purchase commitment.

The Economic Case:
Rest Rescue transforms a defensive cost (preventing churn) into offensive value creation:

Traditional Approach:

  • Ignore declining engagement until churn
  • Attempt expensive reacquisition (5-10x retention cost)
  • Success rate: 10-20%
  • Waste: 80-90% of reacquisition spend

Rest Rescue Approach:

  • Intervene during Best→Rest transition
  • Systematic, scaled, automated intervention
  • Success rate: 40-60% (intervening early)
  • Cost: 10-20% of reacquisition spend

Rest Rescue doesn’t eliminate all churn—some customers leave for legitimate reasons (moved, changed needs, competitive preference). But it eliminates preventable churn caused by inattention, ensuring that every customer who could be saved, is saved, at the optimal intervention point.