Published July 27, 2026
1
Marketing’s Missing Middle
Strip marketing down to its essentials and most brands are running exactly two engines. CRM handles retention: app, on-site personalisation, email, WhatsApp — all owned channels, all cheap, all effective for exactly as long as the customer keeps paying attention. Adtech handles acquisition: Google, Meta, marketplaces — powerful because it can find someone even after the brand’s owned channels have gone silent, expensive because that power is rented, not owned. For a genuinely new customer, that rent may be unavoidable. For a customer the brand already knows, it has a harsher name: AdWaste — paying a platform to reach a relationship the brand had already earned.
Two engines, doing two different jobs, is a perfectly reasonable way to run marketing — until you notice what happens in between them. A customer doesn’t move directly from engaged to lost. There’s a long middle stretch where something stalls: a KYC remains incomplete, a lead goes quiet, a quote sits unclosed, a renewal hangs pending, a first buyer never repeats, a Best customer begins to drift. CRM usually tried — a reminder went out, a journey fired — and then, because that’s how campaign tooling is built, it stopped trying. At that point most brands do one of four things: escalate to a call centre, send a list to an agency, push the customer into retargeting, or nothing at all until adtech sells the customer back. The brand ends up buying back a relationship it already had.
The fix isn’t a third engine. It’s making the two you have refuse to let anyone fall through silently.
NeoMarketing doesn’t propose scrapping CRM and adtech for something new. It keeps both engines and gives each one a second, accountable zone — a place to catch exactly what the original two-engine model was built to drop.
The retention engine is called Meridian. It exists for Best customers — the ones already delivering outsized value — and its job is maximising their lifetime value. Meridian runs two zones. Its normal zone is Retain: CRM 2.0, agentic marketing running on modern channels, doing what good retention has always tried to do. Its stress zone is Finish — the zone that didn’t used to exist. When a Best customer’s renewal quote goes stale, or their KYC upgrade stalls halfway through, Finish — delivered by Progency — completes the outcome rather than letting it quietly die in a queue. Both zones answer to the same discipline: Never Lose Customers.
The acquisition engine is called Atrium. It exists for Rest and Next customers — the ones who’ve gone quiet, and the ones never yet acquired — and its job is pushing the cost of reaching them toward zero. Atrium also runs two zones. Its recovery zone is Recover: Progency reactivating a dormant customer on owned channels before the brand ever reaches for its ad budget to buy them back. Its acquisition zone is Acquire: NeoNet, powered by ActionAds, finding new customers cooperatively — through attention another brand in the network has already earned — rather than renting cold attention from a platform that has none. Both zones answer to a different discipline: Never Pay Twice.
The third NEVER — Never Pay Fixed — is deliberately not a fifth zone. It’s the pricing discipline running underneath the accountable work: Beta as the baseline, Alpha as the verified uplift, Carry as the share of upside, applied through Finish and Recover and eventually through parts of NeoNet. In plain language: NeoMarketing doesn’t ask a brand to pay for activity. It asks to be paid for outcomes, proven against what would have happened anyway.
Notice the pattern in the diagram: Finish and Recover — the two zones that didn’t exist in ordinary marketing — are both delivered by the same team, Progency, even though they sit inside two different engines. That’s deliberate, and part two explains why it matters more than it first appears.
One limit worth stating honestly, up front rather than in a footnote. NeoNet reduces dependence on adtech; it doesn’t eliminate every form of paid acquisition. It works when someone in the cooperative network already knows the customer. Truly cold, net-new-to-category acquisition may still require paid media. The goal isn’t to abolish adtech overnight. The goal is to make adtech the fallback, not the first reflex.
NeoMarketing is the anti-martech, zero-AdWaste operating system for customer value: Retain what is active, Finish what is stuck, Recover what is dormant, and Acquire through trusted attention before renting reach from platforms.
The four zones at a glance
| Engine | Zone | Offer | Job | NEVER |
| Meridian | Retain | CRM 2.0 | Grow and protect active customers | Never Lose Customers |
| Meridian | Finish | Progency-Finish | Complete stuck outcomes after CRM stalls | Never Lose Customers |
| Atrium | Recover | Progency-Recover | Reactivate dormant customers before adtech | Never Pay Twice |
| Atrium | Acquire | NeoNet + ActionAds | Cooperative acquisition through trusted attention | Never Pay Twice |
| Cross-cutting | — | Alpha pricing | Beta + Alpha + Carry | Never Pay Fixed |
2
From Framework to Operating System
Four zones and two engines make a clean diagram. What makes the diagram true rather than just tidy is that each zone is answering a specific, nameable failure — and that there’s a working system underneath, with real pricing discipline, a real build plan, and a real moat. This part covers both: the capabilities that make the zones function, and the operating decisions that keep them honest.
The Three A’s: why each zone actually works
Agentic closes the intelligence gap. Most brands don’t actually know their customers individually — they know segments, and segments are averages. Agentic capability, built on BrandTwins and coordinated through M-Agents, is what makes Retain and Finish possible at the level of a single customer’s specific, current context rather than a cohort’s general tendency. Without it, Retain degenerates into generic campaigns, and Finish has no way to know what, precisely, a given customer left unfinished. This is what makes Never Lose Customers achievable rather than aspirational.
Alpha closes the incentive gap. Knowing what to do isn’t the same as being paid fairly for doing it. Alpha pricing is what lets Finish and Recover be delivered without reintroducing the exact problem they exist to solve: paying for activity instead of results. Every intervention in these two zones is measured against a randomised, concurrent holdout — the brand’s current best effort, never a fictional “no contact” world — so the brand pays only for lift it can independently verify. This is Never Pay Fixed, made structural rather than promised.
Attention closes the attention gap. A customer can be technically reachable and still be drifting, because nothing in their inbox has given them a reason to keep opening. NeoMails, with habit-forming Magnets and the currency Mu, exists to keep a relationship warm independent of any transaction — the “Relate” category most brands send zero of today. This is what makes Recover and Acquire genuinely different from paying adtech twice for the same person: the attention being spent is owned and earned, not rented at a CPM. This is Never Pay Twice, operationalised.
Why the coral zones matter: Progency is not a fifth product
The most important detail in the four-zone diagram is the colour coding. The coral zones — Finish and Recover — are both Progency, regardless of which engine they sit inside. Finish sits inside Meridian because it deals with customers whose value is still visible but whose outcome has stalled. Recover sits inside Atrium because it deals with customers whose attention has decayed and must be rebuilt before paid media takes over. One accountable delivery function, stretched across two different customer states.
That framing prevents a failure mode worth naming: Progency becoming a fifth product bolted onto NeoMarketing with its own pitch, its own pipeline, and its own drift. It isn’t. It’s the accountable layer inside the two stress zones — and it operates only on declared leakage pools: named customer groups where the brand agrees, upfront, that value is leaking, the current best effort is insufficient, and the desired outcome can be measured. Progency doesn’t run BAU CRM. The in-house team keeps everything that’s working; Progency takes what’s stuck.
Sequencing: Finish first, Recover close behind
Finish should usually be the first proof path. Its cycles are shorter and its baselines cleaner: incomplete KYC, stuck leads, unclosed quotes, renewal completion, form completion, Pay-in-Email for active buyers, usable data capture. The customer is still warm enough to act, and the brand usually knows the unfinished job. The alternative being displaced is a call centre, a manual queue, an agency — or nothing — which means pricing anchors to the value of completed outcomes or the brand’s current cost per completion, not automatically to adtech.
Recover is the bigger strategic prize, and the harder one. Dormant customers have lower attention, staler signals, and more uncertain intent. But the alternative being displaced is usually adtech reacquisition, which makes the economic story sharper: recover this customer on owned channels before paying Google, Meta, or a marketplace to win back someone the brand already knew. The proof standard doesn’t change — treatment versus current best effort, with a holdout, and Carry only on verified Alpha.
One pricing temptation deserves explicit discipline: a flat price per email open. It looks like an easy entry point, and it quietly becomes the same exposure-based mechanic NeoMarketing argues against — a CPM by another name. Weak intent can, at most, sit inside Beta as a small activation floor. The real payment must come from what that intent produces: action, usable data, progress, revenue, or verified uplift. That’s how Never Pay Fixed stays intact under commercial pressure, not just in the doctrine document.
The build: an intelligence layer, not a new platform
The operating stack follows the same discipline. NeoMarketing should not rebuild customer engagement infrastructure from scratch. The right build is an intelligence and operating layer over rails that already exist: CE, email, WhatsApp, RCS, CDP integrations, tracking middleware, Pay-in-Email, and NeoNet and ActionAds where relevant. M-Agents automate the repeatable parts of the runbook; Martech Growth Engineers supply judgement and hold the client relationship. Automation follows repeated runbooks — it is not a prerequisite for the first pilots. The factory is the destination, not the starting requirement.
And the moat is not integration complexity — integration is merely the cost of building. The moat is the Decision Trace Graph: a structured memory of context, action, channel, treatment, holdout status, cost, outcome, and next state, written back from every Finish and Recover engagement. Over time those traces become a proprietary learning system — what actually moves customers across zones, categories, and brands. Models will commoditise. Verified decision memory will not.
The redefinition, in one breath
Two engines, four zones, one Alpha discipline. Meridian maximises LTV through Retain and Finish. Atrium minimises CAC through Recover and Acquire. Alpha pricing ensures the accountable middle is paid on outcomes, not activity — and every intervention writes back a trace that makes the next one sharper. Old marketing measured what went out. NeoMarketing measures what moved, and who’s accountable if it doesn’t.
Retain. Finish. Recover. Acquire.
Max LTV. Minimise CAC. Eliminate AdWaste.
3
Mapping to the TAT
Two Engines, Four Zones is the story a CMO hears first. But it sits on top of an older, more granular instrument: the Transaction-Attention Table, or TAT — the grid that maps every identified customer by how much they’ve bought and whether they’re still paying attention. Rows are transaction depth: None, One, Repeat. Columns are attention recency: Strong (engaged within 30 days), Weakening (31 to 90 days), Lost (90-plus days of silence). Every customer lands in exactly one of nine cells, and each column already carries a verb of its own: Grow the Strong, Protect the Weakening, Recover the Lost.
The four zones and the three TAT columns are describing the same territory at two different altitudes, and leaving the relationship implicit invites a vocabulary split: one team talking in zones, another in columns, both right and unable to tell. TAT is the diagnostic grid a brand’s own analyst builds and reads. NeoMarketing is the accountable operating system built around what that grid shows — it takes the diagnosis and asks who will act, what they’ll be paid for, and how the movement gets proven. Getting the correspondence right matters more than it looks, because it settles a question that would otherwise resurface in every deck: is Retain the same thing as Grow? Is Finish just another name for Protect? The honest answers are close, but not identical, and the gap between close and identical is where a pilot’s scope gets drawn.
Retain maps to Grow, but only partly. Grow spans all three depth rows — converting known prospects who haven’t bought (the First play), moving one-time buyers across the second-purchase inflection (Second), and keeping proven repeat buyers buying (Repeat). Most of that is Track 1 work: CRM 2.0, run directly by the brand, sold as leverage rather than underwritten. Retain, in the Meridian sense — Alpha-priced, outcome-underwritten — is specifically the overlay on the Best/Repeat cell, where the stakes are highest and the measurement is cleanest. The rest of Grow stays exactly where it’s always been.
Recover maps to Recover with no qualification needed — this is the one exact match. TAT’s Lost column already carries the name; the cells inside it, R1, R2, R3, are precisely Progency’s Recover territory, prioritised by which row a lapsed customer fell from. The work runs primarily on owned channels, though NeoNet’s cooperative signal can occasionally sharpen it too — the naming system defines NeoNet itself around “deterministic customer recovery,” so a lapsed customer this brand has lost attention on may still be warm to another brand in the network. That’s a supporting mechanism, not the main one; owned channels do most of Recover’s work. (The R stands for Rest, the segment state, not for Recover, the verb applied to it. Two different words, same letter, easy to conflate in a room.)
Finish is the mapping that needs the most care. It’s tempting to write Finish equals Protect and move on, but that overstates the fit. Protect is attention-state language — a customer drifting from Strong toward Lost, regardless of whether anything specific is unfinished. Finish is outcome-state language — a named job left incomplete: a KYC form, a quote, a renewal, a cart. A customer can sit in the Grow column, fully engaged, and still have an abandoned cart from yesterday that needs finishing. So Finish isn’t contained by Protect. What’s true, and worth saying precisely, is that Finish’s richest pools concentrate in Protect — because a weakening customer with a stalled outcome is exactly where a brand’s own current effort is thinnest, and thin baselines are where a holdout can prove the most. That’s the same logic that already told us not to lead with abandoned cart: the Grow-column version of Finish is real, just rarely the best place to start.
Acquire sits outside the grid entirely, by construction. TAT only plots customers a brand has already identified — even its emptiest cell, None, is full of known people: subscribers, registered users, abandoned browsers. Acquire is for the customer this brand hasn’t met yet, found instead through someone else’s earned attention inside the NeoNet cooperative. The moment that customer responds, they stop being outside the table and become a fresh entry in the None row — Acquire’s job was only ever to be the door.
Put together, the four zones tile the entire TAT and the one population it can’t see, with no zone left unaccounted for and no cell double-claimed. TAT measures the portfolio. NeoMarketing assigns accountability. Progency underwrites the two stress zones. That completeness is the real payoff of doing this mapping properly rather than loosely.
