Version control for marketing claims: what Git teaches commercial truth

A head of GTM ships a repositioning, signs off on the new battlecard, and watches it land in a deck. Six weeks later a rep is still quoting the old competitive line on a live call, an AI agent is grounding from the prompt it was forked from two quarters ago, and nobody can say which version is current. The phrase that surfaces in that moment, almost reflexively, is “we need Git for our marketing claims.”

The instinct is correct. The tagline is not.

The Commercial Truth manifesto argues that marketing is the last commercial function without infrastructure — engineering got Git, finance got the ledger, data got lineage, and the claims a company makes about itself never got a substrate (Source: Assay brand canon, §10 Truth Graph manifesto). Reaching for Git is the right reach. What matters is which lesson you take from it: the literal tool, or the four primitives the tool introduced.

The concept, named plainly

“Git for marketing claims” means applying version control to commercial truth — the positioning statements, proof points, competitive talk tracks, and pricing facts a company asserts. Not a metaphor about moving fast; a literal application of four structural primitives that version control made standard.

The unit changes. Git versions files; commercial truth versions claims. A claim — “we are the only platform that does X,” “our pricing starts at Y” — becomes a first-class object with a history, rather than a sentence buried in a doc that is a snapshot the moment it is published.

This matters now because of a substrate shift. AI agents have proliferated across revenue teams in the last two years, and an agent grounding from an unversioned doc amplifies drift faster than any team can chase (the truth gap). When the source has no version, every consumer of it inherits the ambiguity.

Why the literal tool is the wrong import

The note worth holding is the one in Assay’s own canon: avoid claiming to be “Git for marketing” — the analogy is the lesson, not the tagline (Source: Assay founder essay series, Essay 2). There are concrete reasons.

Git versions text files and resolves them by line. A positioning claim is not a line of text; it is a typed object with a source, a confidence level, and downstream dependents. Diffing two prose paragraphs tells you what characters changed, not whether the claim changed or whether the change was approved.

Git also has no native notion of confidence or source authority. In commercial truth, the difference between a human-verified claim and an AI-extracted one is load-bearing — the model’s certainty cannot override a human’s verification (the confident wrong answer). That is a constraint Git was never built to express.

So the import is not the tool. It is the four primitives the tool made ordinary, re-expressed for claims.

The primitives version control introduced

Before Git, engineering teams lived with lost edits, no merge history, no audit trail, and no way to test a change in isolation (Source: Assay founder essay series, Essay 2). Git’s contribution was four primitives that map directly onto commercial claims.

Nodes — claims as first-class objects. Git made a commit a thing you could name and inspect. The equivalent is a typed node: a PositioningStatement, a CompetitiveTalkTrack, a ProofPoint, a PricingTier (Source: Assay brand canon, §10). The claim stops being a sentence and becomes an object with structure (typed knowledge graph for marketing claims).

Commits — every change is versioned. Every node carries a version history, a valid_from, and a valid_until. You can diff this quarter’s positioning against last quarter the way an engineer diffs a release (Source: Assay brand canon, §10). The change is recorded, not whispered in a Slack message that gets forgotten (the 11pm Slack message that killed a deal).

Signatures — every approval is attributed. Git records who authored a commit. Commercial truth goes further: each node carries a source type with an enforced confidence ceiling, so an AI-extracted claim cannot present itself as human-verified (source-type taxonomy for marketing claims). The signature is not decoration; it gates what an agent is allowed to say.

Audit and tests — history is provable, and changes are validated. Git gave teams a full, tamper-resistant history and a place to run tests before merge. The claim equivalent is a hash-chained audit log plus calibration: a claim is “what we say, last validated three days ago, with a credible interval,” not just “what we say” (Source: Assay brand canon, §10). Where the outcome data is not there, the honest value is “not yet,” not a fabricated number.

There is a fifth move with no clean Git analogue, and it is the one that earns the substrate label: cascade. When a node changes, every dependent surface — every deck, battlecard, landing page, and agent prompt that grounds in it — is flagged for refresh (Source: Assay brand canon, §10). Git propagates to files; cascade propagates to everything that quoted the claim (every AI agent needs a source of truth).

A worked example

Take one claim: “We are the only vendor with native EU data residency.” A marketing team writes it once, in a deck.

Under the doc model, that sentence is copied into a battlecard, paraphrased on a pricing page, summarized in three AI agent prompts, and quoted in a proposal template. Five surfaces, five forks, no link between them. When a competitor ships data residency next month, someone has to remember all five — and the AI prompts, which nobody owns, keep asserting the old claim into live conversations (5 AI tools, 5 versions of your company).

Under the version-control model, the claim is one node. It has a source type (human-verified, from a signed compliance attestation), a version, and a confidence ceiling. The five surfaces do not copy the sentence; they reference the node.

When the competitive landscape changes, an operator edits one node. The source type forces a re-verification step before the change commits. The cascade flags all five dependents the same day, and the audit log records who changed what, when, and on what evidence — which is the difference between a claim you can defend in a board review and one you cannot (what CRM did for pipeline, commercial truth does for knowledge).

The exact reduction in propagation lag this produces across a large org is not yet measured publicly; it is reported as a credible interval in dogfood data, not a point estimate. What the structure changes is categorical, not incremental: the claim went from five unlinked snapshots to one versioned object with a provable history.

What this is not

A system that lacks these primitives is not infrastructure. It is a wiki with ambitions (Source: Assay founder essay series, Essay 2). A shared doc, a tidy Confluence space, a Notion database of approved messaging — each is a better container, and none of them is a substrate, because a container cannot enforce a source type, cannot cascade a change, and cannot prove its own history (we don’t have a content problem, we have a truth problem).

That is the line the Git analogy draws. The lesson is the primitives. The tool is not the import, and “Git for marketing” is not the name.

Closes / opens

Closes LSO §F.6 — the parallels cluster, and specifically the question of what “Git for marketing claims” means and why the analogy holds at the primitive level while failing as a tagline.

Opens the next question: if the primitives are nodes, commits, signatures, audit, and cascade, what is the minimal schema that expresses all five for a commercial claim — and how does an operator adopt it without a full re-platforming? That is the implementation question, and it is where the infrastructure layer argument turns from concept into product.

The methodology Assay is developing for the Commercial Truth Index measures exactly the capacity these primitives create: whether a company’s commercial claims are grounded, calibrated, coherent, and auditable across every surface.

This essay is grounded in the Assay founder essay series (Essay 2, “The Git lesson”) and the Truth Graph manifesto in brand canon §10. Methodology for the Commercial Truth Index is in development.