Published June 20, 2026
How a CRM team gets to Alpha by compressing transaction time and cutting transaction tax — three plays on the grid, three on its edges.
1
The Two Levers — Time and Tax
For two decades, marketing’s unit of work has been the campaign: build a list, send to the list, report what was opened and clicked. That model is quietly failing, and the dashboards that run it cannot see why. Acquisition costs rise every year. Repeat revenue leaks into paid media, where the brand pays a platform tax to win back customers it already owned. The most valuable customers drift silently — purchase intervals stretch, engagement softens — and the P&L only learns about it months later, when the customer reappears as a retargeting line item.
This playbook describes the operating system that replaces the campaign model. Its unit of work is not the campaign; it is the customer’s movement through a grid — and the Alpha that movement produces above a baseline. There is one diagnostic instrument, six plays, and four key numbers a CMO needs to track.
**
Alpha is incremental contribution above a Beta baseline — the revenue a brand would have earned anyway. It comes from exactly two places, and a move that touches neither is not Alpha, however busy it looks. The first lever is time: how long a customer takes to do the next thing. The second is tax: how much it costs to make that thing happen. Everything in this playbook is one of those two levers pulled deliberately.
Take time first. Transactions take time because customers drift, and drift is invisible until it is expensive. A customer who used to buy every six weeks slips to ten, then twelve, then stops opening messages altogether. Each stretch is margin foregone and a step closer to the retargeting pool. The time lever has two halves: speed the build — move a customer from Next to Test to Best faster — and slow the drift — delay the slide from Best to weakening to Lost. Compress the gaps on the way up; defend against the gaps on the way down.
Now tax. Every transaction reaches the customer by a route, and every route carries a tax. Organic costs almost nothing. Owned CRM is cheap on paper — five to ten per cent — until you add the discounts, coupons and loyalty burn that bought the sale, at which point a brand’s ‘cheap’ revenue can be running adtech-like economics through its own list. Paid media takes a fifth of revenue; marketplaces take more than a third. And the most expensive tax of all is the one no dashboard names: paying paid media to re-buy a customer the brand already owned. The tax lever means moving revenue down that ladder — fewer transactions on the expensive rungs, more on the cheap ones.
These two levers are the whole game. Less time, less tax, measured against a baseline, equals Alpha. The rest of the playbook is the instrument that shows where the time and tax are leaking, and the six plays that pull the two levers in the places it leaks most.
2
TAT: The Grid to Build
Standard segmentation is recency, frequency, monetary value — RFM. It looks backward, because it is built from transactions alone. The Transactions-Attention Table (TAT) looks forward, because it adds the one axis RFM ignores: attention. Attention is the signal that tells you a customer is drifting before the revenue loss appears. The TAT crosses it with transaction depth, and every customer lands in exactly one cell.
Rows are depth: zero orders, one to two, three or more. Columns are attention recency: strong (engaged in the last 30 days), weakening (30 to 90), lost (90 days or more). Both axes run to eternity, and that is the design choice that matters. Transaction depth counts a customer’s whole history; the ‘lost’ column has no upper bound. So the customer who has never bought and has not engaged in a year is not some separate ‘dormant’ category that lives off the grid — it is simply the top-right cell, zero transactions and attention lost.
That corner cell is usually the largest on the grid, and naming it honestly is the first useful shock. Most of what a brand calls its ‘identified base’ is people who never bought and are no longer listening. That number is a vanity metric; the reachable base is a fraction of it. The corner is a cost-and-compliance question — repermission a scored slice, suppress the rest — not a place to spend Alpha. Putting it on the grid rather than in a footnote keeps a CMO from mistaking list size for reach.
Nine cells are right for an analyst and too many for a first conversation, so they collapse to four. Do they buy, by do they still pay attention.
The first cut is buyers versus prospects, not best versus the rest, and the reason is diagnostic. That single cut is what separates a retention problem from a conversion problem. A brand whose grid is mostly buyers who are drifting has a retention disease; a brand whose grid is mostly prospects who never convert has a conversion disease. The same instrument reads both. The four-cell view flags which one a brand has; the nine-cell view tells you whether a drifting cohort is cheap to protect or expensive to recover.
3
Grow, Protect, Recover: The 3 Grid Plays
The three attention columns are not just states; they are the three actions a brand can take, and they already have names: sell when attention is strong (Grow), relate when it is weakening (Protect), recover when it is lost (Recover). So the three Grid Plays map cleanly onto the three columns. That alignment is the spine of the whole playbook.
Grow runs the strong column. When a customer is engaged, the job is to transact with them on an owned route — and to do it before paid media claims the credit. That means three things at three depths: convert an engaged prospect to a first purchase, accelerate a one-or-two-time buyer to the next, and own the repeat of a proven buyer rather than letting retargeting re-sell them their own intention. Grow is where the time lever and the tax lever meet — it compresses the journey and keeps the transaction off the expensive rungs. It carries two numbers, one per lever — Time-to-Next-Transaction for the time it compresses, Paid Repeat Leakage for the tax it cuts — and it serves Never Pay Twice.
| Depth | Grow’s job at that depth |
| 0 orders · Next | Convert the first purchase on an owned route, while attention is live |
| 1–2 orders · Test | Accelerate the next purchase; compress the gap between orders |
| 3+ orders · Best | Own the repeat; keep it off paid media so adtech cannot re-sell the customer their own intention |
Protect runs the weakening column. These are customers with purchase proof whose attention is fading — the cheapest intervention on the entire grid, because catching a customer here costs a fraction of recovering them once they are gone. The play is counter-intuitive: pause the promotional pressure, lead with utility, recognition and service, and restore relevance before the relationship breaks. Its number is the Weakening Pool (proven buyers in the weakening column = B⁻ + T⁻, the Best and Test who are drifting), and it serves Never Lose Customers. Most brands do the opposite — they discount harder at a drifting customer, accelerating exactly the loss they fear.
Recover runs the lost column. Attention is gone, so the sequence is attention first, transaction second — restore engagement before asking for a sale, and do it before defaulting to paid retargeting. Its number is Rest Recoverable Value: the historical revenue sitting in former Best and Test customers who have gone dark. It serves Never Pay Twice, because every customer recovered through restored attention is one not re-bought at full adtech price.
The column tells you which play to run; the row tells you how hard to run it. A Best customer who is weakening earns more protection investment than a never-buyer who is drifting. Attention sets the action; depth sets the intensity. That is the entire logic of the Grid Plays.
4
Capture, Acquire, Monetise: The three Edge Plays
The three Grid Plays act inside the grid, moving and holding customers by column. The three Edge Plays are different in kind: they act on the grid’s edges. They bring customers onto the grid, or they extract value off it. That is why they are Edge Plays — not because they matter less, but because they operate at the boundary rather than within the lifecycle.
Capture Identity brings buyers onto the grid as known customers, and it has two surfaces. The first is the intermediated buyer — a marketplace customer who is platform-owned for life until the brand captures a direct identity. The second is the anonymous visitor — traffic, often paid-acquired, that lands on the brand’s own site and leaves unidentified; for a D2C brand that second leak is often the larger one. The play adds value-exchange capture before checkout and post-purchase capture afterwards — a sample for an email, a QR code on the packaging, a warranty that requires registration, a reason to log in. Its number is Identity Capture Rate, and for a marketplace-heavy or paid-traffic-heavy brand it is the single most consequential play in the book. That is the point worth stressing: an Edge Play is named for where it acts, not for how much it matters. For some brands the most urgent move is an Edge Play.
Acquire brings genuinely new identified customers onto the grid without paying the full adtech tax. This is NeoNet — cooperative recovery and acquisition across brand boundaries, where one brand’s lapsed customer is another’s warm prospect. It is the only honest way to grow the base that does not route every new customer through a platform that will then charge the brand to reach them again. Concretely: a premium skincare brand can pick up a high-intent beauty buyer through a complementary fashion brand’s NeoMail — a one-tap sample request, at a fixed transfer fee — instead of renting that same buyer back through Meta at auction price.
Monetise Attention extracts revenue from attention that is not transacting. The drifting and lost cells still open some emails; that attention has value even when the customer is not buying. ActionAds across owned surfaces turn it into revenue, which funds the recovery layer that works on those same cells. It is the one play that pays for the others. Run as a structural test first — its economics are unproven in some categories — but the principle is sound: attention is an asset whether or not it transacts today.
5
Who Runs Them — CRM 1.0, CRM 2.0 and Progency
A playbook that cannot say who does the work on Monday is a philosophy, not a plan. Almost all of it stays with the brand’s own people. The CRM team runs Grow and Protect; the acquisition team runs Acquire. Only one play leaves the building — and only one. Progency, the outcome-priced recovery business, takes the lost column and nothing else. It is deliberately post-CRM and pre-Adtech: it works the gap the CRM team cannot reach and stops the brand falling through to paid media, where the reacquisition waste happens. It is an aid to the CRM team with skin in the game, not a threat to anyone’s job.
Start with what the CRM team does today, because most brands are here. CRM 1.0 is a marketing automation platform, staffed by the vendor’s customer-success managers, pushing batch campaigns through owned channels, measured on opens and clicks, paid for as a fixed software fee. Its unit of work is the campaign sent. It is competent at delivery and blind to movement; it can tell you what was opened but not which customers are weakening before the revenue falls.
CRM 2.0 is the same team with a better operating system. The marketing automation platform is enhanced with M-Agents (Insights, Audience, Content, Decisioning). The vendor’s customer-success managers become Martech Growth Engineers with proactive assistance; batch channels become interactive Channels 2.0; the engine underneath is the shared Context Graph; and the measure changes from opens and clicks to velocity, Real Reach, and Click Retention Rate (CRR). The economics shift from a flat fee to a fee plus upside, because the work is now accountable for movement, not activity. This is what Agentic Marketing offers the in-house team: not a rip-and-replace, but a re-pointing of the same people at the two levers.
Two of those in-house plays lean on a shared engine designed to pay for itself. Atrium is the attention engine — NeoMails as the Relate vehicle (today email, extensible to WhatsApp and beyond), ActionAds as the monetising unit inside them, and NeoNet as the network that places those ActionAds. Its design goal is to be self-funding: at scale, the ActionAds pay for the NeoMails on a ZeroCPM basis and return the residual to the brand, so Protect approaches zero operating cost and Acquire — through NeoNet — brings new identified customers without the full adtech tax. The early-pilot test is narrower: whether attention yield can cover a meaningful share of sending and content costs. The ambition is that the brand eventually writes no cheque for Atrium because the attention pays for itself — stated as a target the pilots must prove, not a fact assumed.
Progency is the one piece the brand buys as an outcome rather than runs as a capability, and the question it must answer is blunt: how will it recover lost customers better than the CRM team could? The answer is two engines working as one inside a black box. Atrium earns the attention — the same self-funding NeoMails, now aimed at the lost. Meridian converts it: a proprietary model with an Alpha Agent that turns earned attention into transactions — the Alpha Agent being the artifact that has internalised the judgement of the best CRM operators and MGEs, built up over time. Meridian is never sold and never shown, the way a quant fund’s model is the secret it is paid for, not a product it ships; what ships are the M-Agents inside CRM 2.0. And it is priced purely on what it provably delivers: a share of the incremental transactions it converts, measured against a held-out control and benchmarked at a fraction of what adtech charges to reacquire the same customer — no fixed fee. The test of whether a partner believes its own numbers: ask whether it will hold back a control group and take its fee only on the lift.
| CRM 1.0 · today | CRM 2.0 · brand team | Progency · done-for-you | |
| Support | CSMs (vendor) | MGEs (vendor) | MGEs + Alpha Agent |
| Stack | Channels 1.0 + Martech | Channels 2.0 + M-Agents | Atrium + Meridian |
| Engine | Rules and segments | Context Graphs | BrandTwins, TwinFactory |
| Focus | Campaigns sent | Next → Test → Best velocity | Recover the lost |
| Measure | Opens, clicks | Velocity, Real Reach | Alpha Generated |
| Economics | Fixed SaaS fee | Fixed + upside | Share of proven lift |
| The job | Sends campaigns | Moves customers | Owns the outcome |
6
The Numbers that Prove It
A framework that cannot be measured cannot be defended, and this one rests on three layers of number. The first layer is the four diagnostic numbers, each of which triggers exactly one play.
Paid Repeat Leakage is the share of repeat revenue bought back through paid media; it triggers Grow. The Weakening Pool is the count of proven buyers losing attention; it triggers Protect. Rest Recoverable Value is the historical revenue in lost Best and Test customers; it triggers Recover. Identity Capture Rate is the share of intermediated buyers and anonymous visitors converted to known customers; it triggers Capture. Four numbers, surfaced on the first Alpha Audit, that turn a vague sense of ‘marketing could be better’ into a board-ready slide titled ‘how much of last quarter’s spend was structurally avoidable.’
The second layer is the governance number: Alpha above Beta. Beta is the revenue trajectory the brand would have earned anyway; Alpha is the incremental contribution the plays produce above it. The discipline that keeps Alpha honest is measurement against a pre-agreed baseline with matched controls — not Alpha against zero, which is the oldest deception in marketing. Every play in this book is run as a test against a held-out cohort, and only the lift that survives the control is counted.
The third layer is the number that actually moves a budget, and it belongs to the CFO. It is the ratio between what a Progency-recovered dollar returns and what an adtech dollar returns — because recovery and adtech compete for the very same reacquisition budget.
Run Progency to roughly twice the return of adtech on the same lapsed customers — a ten against a five — and every reacquisition dollar moved from the paid channel to recovery creates net value. The claim is not asserted but measured: the three-arm holdout puts Progency and adtech against one shared control and reports the cost per incremental recovery for each. The gap is durable and widening — adtech returns erode under auction inflation and signal loss, while recovery improves through restored attention, lower offer tax, and the suppression that stops paid media claiming credit for customers who would have returned anyway. It survives a sceptical CFO only if it is net of offer tax, incrementally proven against the control, fully costed, and framed as a measured result rather than a promised multiple. And the sequence is non-negotiable: exhaust the owned channel first, then recover through Progency, and only then spend adtech — the reacquisition dollar should reach the cheapest proven route before the most expensive one.
7
Getting Started — the Data and the First Move
The system begins not with a campaign but with a classification, and that needs data the marketing team rarely holds alone. The first job of the CMO is to convene finance, commerce and engineering for the data the audit requires — and if that convening is impossible, that is itself the first finding.
| Data source | Fields required | Likely owner |
| Orders | Order ID, customer ID, date, revenue, channel, platform | Commerce / Data |
| Customer master | First order date, lifetime orders, lifetime revenue, contactability | CRM / CDP |
| Paid media | Campaign, spend, prospecting vs retargeting, customer match | Growth |
| CRM events | Opens, clicks, push taps, WhatsApp reads, app sessions, timestamps | CRM / Martech |
| Offers | Coupon, discount, cashback, free shipping, loyalty burn | Finance / CRM |
| Intermediaries | Commission, discounts, identity availability | Marketplace |
With the data joined, the first thirty days are diagnostic, and they compress to five steps. Build the transaction file. Classify every transaction by route and tax. Compute the effective tax, route plus offer, so the cheap channels stop pretending to be cheap. Build the TAT by joining each customer’s last attention event to their transaction history. Surface the four numbers. Perfection is not the goal at this stage; consistency is. Lock the attribution rule and the attention-event definition for a quarter, document them, and refuse to relitigate them every time a number looks ugly.
Then run one play, not six. Start with the half of Grow that owns the repeat: suppress your active customers from retargeting and redirect the saved spend to owned channels. It is the fastest result in the book — the mechanic is mechanical, remove an audience and measure the lift, with first signal in 30 to 45 days — and it pays for the entire Alpha Audit. It is also the safest place to start a negotiation with a sceptical performance team, because it settles the argument with a suppression test rather than an opinion. Do not start with campaigns. Start with classification, then the one play that pays for itself, and let the proof earn the right to the next five.
That is the whole playbook on paper — two levers, one grid, six plays, four numbers, and a first move that funds the rest; what it looks like in a real quarter is a story best told through the person who has to run it.
8
Maya’s journey from Beta to Alpha
Maya had run marketing at the skincare brand for three years, and by every measure on her dashboard she was winning. Revenue was up, acquisition was scaling, the campaigns shipped on time. The board nodded each quarter. And yet the contribution margin kept thinning, and she could not say why.
The Alpha Audit took three weeks and the numbers landed like a diagnosis. Thirty-eight per cent of her repeat revenue was being bought back through paid media — she had assumed it was under fifteen. Her ‘cheap’ CRM revenue, once she added the discounts, was running at an effective tax near a quarter. Half her identified base had not engaged in a year. And a fat band of her best customers — the ones who had bought three times or more — were sitting in the weakening column, drifting, months before any transaction report would have caught them.
The hardest conversation came first, and it nearly stopped her. ‘If you pull my actives out of retargeting, revenue will fall — and it will be on you,’ her performance lead said. Maya did not argue the claim. ‘Then the control will prove it,’ she said. Ninety-day-active customers were pulled from the retargeting audience, the saved budget redirected to owned reactivation, and — the move that made it safe — a matched group was held back, untouched, as the measuring stick. Thirty-five days later the answer was on the table: the held-out group came back almost as often as the retargeted one. Most of that ‘incremental’ spend had been buying customers who would have returned anyway. The performance lead went quiet; the number was his own. That was Grow, owning the repeat, and it paid for the whole audit by the end of the quarter.
She did not stop there. She ran Protect on the drifting Best cohort — paused the discounts that were training them to wait, replaced them with relationship messaging — and watched the open rates stabilise first, the transactions follow at sixty days. For the lost column she did not staff a team; she handed it to Progency, on a baseline-plus-Alpha arrangement with no fixed fee, and let the recovery engines do work her CRM team was never built for. The marketplace identity problem went into a capture programme: a QR code on the carton, a reason to register.
At the ninety-day board review, Maya presented one number she had never had before. Not revenue, not ROAS — Alpha, measured against an agreed baseline, with the controls that proved it was real. The CFO, who had spent the quarter sceptical, asked the only question that mattered: if Progency delivers twice what an adtech dollar returns, why are we still spending the adtech dollar? Maya had the sequencing answer ready — CRM, Progency, and then Adtech.
What changed was not Maya’s effort. It was her unit of work. She had stopped running campaigns and started moving customers, and she could prove what the movement was worth. Her old job had been to send more; her new job was to lose less, own more, and recover before paying twice. The thinning margin the P&L had felt for years finally had a name on the dashboard — and three rules to fix it: Never Lose Customers. Never Pay Twice. Never Pay Fixed. That became Maya’s mantra.
***
NeoMarketing creates Alpha by compressing customer time, cutting transaction tax, and recovering customers before adtech makes the brand pay twice. Three Grid Plays — Grow, Protect, Recover. Three Edge Plays — Capture, Acquire, Monetise. Two levers, one North Star Metric: Alpha above Beta. Stop running campaigns; start moving customers.
