The Paper Trail of an Idea

On 9 October 2026, Scott Brinker argued that context has become martech’s word of the year, and that the work behind it is larger than the demos. In “Context” is a hot word in martech, but the work required is underestimated, he separates customer context, company context, and the systems that are supposed to make either one actionable. He names an Alignment Gap and a Coordination Gap that AI speed has opened. His prediction is specific. If context was the word for 2026, 2027 will be the year of coherence.

I have been writing about context, absorption, and the debts that speed creates since August 2025, so the prediction felt familiar. Familiar is not a record. Brinker’s October 2026 prediction makes coherence the word for 2027. The public chain already connected context, absorption, metabolism, and the debts that speed creates, and the dates are what keep that chain from collapsing into the new word.

Brinker is precise about the three contexts, and the precision is the part worth granting. Customer context is the customer’s situation, goals, history, and the moments that matter. Company context is the company’s goals, knowledge, processes, governance, and priorities. Systems context is the harder slice: how much of the first two actually exists in systems in a form that can be acted on. Capturing data, he notes, does not make it actionable. The goal of context engineering, in his account, is to widen that coverage. The goal of value engineering is to match the right company context to the right customer context. He calls the overlap Golden Context, and he treats it as an aspiration in the same family as the old dream of a perfect customer record.

The time-scale picture underneath that model is older than this newsletter. He says he mapped Frank Duffy and Stewart Brand’s shearing layers onto marketing in his 2016 book, Hacking Marketing. Site, structure, skin, services, space plan, and stuff change at different speeds. Slow layers remember and constrain. Fast layers learn and absorb shocks. The trouble shows up at the seams. In marketing, his layers run from the company and the brand down through campaign, channel, tactic, iteration, and the feedback from a single customer. A real-time offer happens inside a campaign, which happens inside a brand, which happens inside a business. AI, in this telling, is an earthquake at those seams.

The Alignment Gap is what happens when the old brake disappears. Launching a campaign used to take enough people, and enough time, that someone experienced could still say the work had drifted. Brinker does not romanticize the slowness. He does say it created room for strategic context to be considered along the way. AI lets more marketers generate more campaigns, and more specialized ones, from a few sentences. The flood can overflow the levees that kept execution aligned with the brand. His remedy is partly technical. Bring more company context into systems, in a form machines can use, and keep that context current as campaigns, pages, and sequences depend on it. Then he adds the part software cannot finish. Someone still has to decide where a human stays in the loop, who that human is, what authority they have, and what time they are given to use it.

The Coordination Gap sits one layer over. Once campaigns can personalize and then improve themselves, each program can evolve in ways the people nominally in charge cannot fully see. If the guardrails from the alignment problem are in place, he treats that evolution as a feature. The new gap is between programs. One campaign optimizes. Another campaign optimizes. The customer, who meets both, experiences a collection of locally successful moments that do not hold together. Closing it, he says, will take orchestration across those parallel agents: control, coordination, cadence, and calibration, wrapped around company and customer context. He expects the practical architecture in a Martech for 2027 report. Coherence, in his close, is how all of these contexts hold together as a whole. He means a brand.

He also says the work is “way bigger than a couple of .md files and a log of customer interactions.” I am going to come back to that sentence. First, the grant. These gaps are real for the teams he is writing to. I am not arguing that martech invented a phantom, or that a brand can ignore what happens when a hundred generated sequences learn on their own.

Two gaps, two claims

Proximity still has a sequence, and two similarly named problems are not one problem.

His Coordination Gap lives between campaigns. Each one learns from its own results. The customer crosses them and meets a fragment. Coordination Debt, which I published on 13 July 2026, lives somewhere else. AI tools multiply briefs, analyses, and handoffs faster than anyone redesigns who can commit, who reviews, and who can override when two workstreams collide. A generated campaign can look finished. It still creates a demand for judgment. Local speed produces unfinished decisions, conflicting authorities, and outputs no one owns. The customer may never see that collision. The institution does, in the meeting that cannot close.

His markdown sentence meets a distinction I published on 22 June 2026 in Knowledge as Code. A portable file can be the substrate. It can carry metadata, and it can carry domain meaning. It does not supply the system: which sources deserve trust, which claims have expired, and which rules govern their use. A company can write down everything it knows and still hand an agent a pile of files with no account of what to encounter first. Knowledge Priming, published on 20 July 2026, named that next question as an attention budget. Who decides what the agent sees before it sees the average of everything else?

The neighborhood was already inhabited. Absorptive capacity has a long academic history. Knowledge management, organizational learning, and domain-driven design all shaped the vocabulary I use. Brinker is not lifting a sealed idea out of this archive, and I am not lifting his pace layers out of a 2016 book. Several people can watch the same acceleration and arrive at nearby conclusions. What the conclusions cannot survive is a scrambled order. Names matter because readers use them as handles. Dates matter because handles get reused. The order in which the handles connect is the part a later essay can erase without meaning to.

What the dates actually say

I went into the archive expecting the context essay to sit first. In late August 2025 I published Mastering Context for Continuous Learning and Market Success, and I have told the story as if that piece opened the sequence. The argument there starts in product marketing. A product, a message, or a strategy becomes intelligible through the frame around it. Buyers need that frame to compare. Teams need it to learn. I called one of the shared maps a context graph: personas, stages, objections, and signals, so each channel would reinforce one account instead of emitting another isolated message.

The folder disagrees with that memory by six days. The Digestion Gap went out on 23 August 2025. Mastering Context went out on 29 August. The digestion essay named the widening distance between the volume of change entering an organization and its capacity to integrate that change into judgment, behavior, and memory. The context essay, the same month, named a marketing version of a shared map. If a private note someday shows that the thinking ran the other direction, that note is a different trail. The public one starts with digestion.

On 25 October 2025, Diagnosing the Modern Organization named four forces behind the gap. Speed Inflation is what happens when AI compresses work to machine tempo while hierarchies still move at human tempo. Decisions start arriving before understanding. Context Collapse is what happens when tools and outputs multiply faster than a team can keep a shared account of what the work means. Each group grows its own language and its own local picture. Activity rises. Collective sensemaking thins out. The same essay names Cognitive Debt and the Trust Deficit, and it contains the word I am now careful about. A company can be rich in tools and, in that essay’s phrase, starved of coherence.

Designing for Absorption, on 17 November 2025, turned the diagnosis into an architectural question. Organizational Metabolism is the rate at which an institution converts change into lasting capability. I have since described the desired outcome of that essay as coherence. The word is not in it. The goal sentence arrives later. On 1 January 2026, A Year of Integration said the goal should be coherence, not velocity. On 23 March 2026, The Leadership Primitive said it outright. The goal is no longer just velocity. It is coherence.

Knowledge as Code followed on 22 June. Coordination Debt followed on 13 July. Knowledge Priming followed on 20 July. I had been telling the July story with priming first, as if the attention budget were the step that made the debt visible. The dates run the other way. The debt essay is the one that asks who can commit. The priming essay is the one that asks what the agent should encounter first.

Read in publication order, the chain is ordinary and checkable. Context connects to absorption. Absorption connects to metabolism. Acceleration shows up as Speed Inflation and Context Collapse. Those pressures accumulate as Cognitive Debt and Coordination Debt. Knowledge as Code and Knowledge Priming are part of the infrastructure for keeping the account trustworthy. Coherence is the organizational condition that chain is trying to preserve. The word has a shorter public life than the architecture around it, and a later prediction can still be right about the year without being first about the problem.

What the record is for

Writers with large platforms often become the moment an idea enters a wider conversation. Their reach gives the concept a second life. Earlier work can fade, even when it held a fuller set of connections, because a new audience meets the idea at the point of arrival.

The words can stay common property. Context, coordination, and coherence do not belong to a newsletter or to a martech forecast. What I can keep legible is the system those words started naming in public here: why faster production can weaken organizational learning, why a file is not yet a governed account, and why the capacity to absorb change will shape the next phase of AI adoption. If other people are reaching similar conclusions, the field may be arriving at the problem. The useful response is to keep writing, keep dating the work, and keep the architecture visible enough that a later word cannot swallow the chain.

Which idea in your own work deserves a clearer paper trail before the wider conversation forgets how long you have been building it?

Madam I’m Adam

This continues a public sequence that runs from The Digestion Gap through Organizational Metabolism, Knowledge as Code, and Coordination Debt. The dates are the thread.