ChapterCH03 Chapter 3 of 12

The GTM Growth System

How does the machinery work, and who decides?

The GTM Growth System

ModelM01Model

The canonical model is Sense → Understand → Decide → Activate → Measure → Learn → Optimise. Each stage has a human owner or an owned service. A loop closes when outcomes inform a later decision; it does not require autonomous execution.

Stage	What happens	Minimum evidence or output	Economic purpose
Sense	Detect customer, market or operating change	Dated signal and source	Recognise an opportunity or avoidable loss
Understand	Check facts, needs and constraints	Diagnosis, alternatives and uncertainty	Avoid solving the wrong problem
Decide	Select a feasible action or no change	Named owner, rationale, limits	Allocate money and capacity
Activate	Deliver the approved intervention	Action version, delivery or receipt	Change a customer outcome or work cost
Measure	Compare outcomes and reconcile expense	Scoped estimate and guardrails	Determine commercial consequence
Learn	Adjudicate what is reusable	Finding, limits, owner and review trigger	Reduce repeat error and investigation
Optimise	Repeat, revise, expand or stop	Next decision and revised Play	Improve subsequent economics

Different decisions use different clocks. A stock or contact-permission check may be time critical. Cohort review may be monthly; positioning changes less often. Faster cycling can contaminate experiments and overreact to noise. Use the required decision latency, not “real time” as a general aspiration.

Sense → Understand → Decide → Activate → Measure → Learn → Optimise. Each stage has a human owner or an owned service.

Text version

Each stage has a human owner or an owned service.

Sense — What happens: Detect customer, market or operating change. Minimum evidence or output: Dated signal and source.
Understand — What happens: Check facts, needs and constraints. Minimum evidence or output: Diagnosis, alternatives and uncertainty.
Decide — What happens: Select a feasible action or no change. Minimum evidence or output: Named owner, rationale, limits.
Activate — What happens: Deliver the approved intervention. Minimum evidence or output: Action version, delivery or receipt.
Measure — What happens: Compare outcomes and reconcile expense. Minimum evidence or output: Scoped estimate and guardrails.
Learn — What happens: Adjudicate what is reusable. Minimum evidence or output: Finding, limits, owner and review trigger.
Optimise — What happens: Repeat, revise, expand or stop. Minimum evidence or output: Next decision and revised Play.
A loop closes when outcomes inform a later decision; it does not require autonomous execution.

Seven logical responsibilities; not seven products. Use this model with the conditions and limits described above.

Decisions and methods

Decision	Evidence	Human owner	Method and AI role	Action and measurement
Which market or proposition?	Customer needs, demand, competition, delivery and economics	GTM/Product leader	Human judgement; optional source synthesis	Approved proposition or market test; contribution and strategic coverage
Which customer intervention?	Eligibility, need, service history, treatment evidence	Lifecycle/Service owner	Rules first; validated models if useful	Repair, guidance, offer or no contact; net assigned-population contribution
Which acquisition signal?	Cohort value and outcome quality	Performance owner with Finance	Deterministic joins; optional calibrated prediction	Native value feed; incremental cohort contribution
How much investment?	Tranche effects, costs, constraints and uncertainty	Budget owner with required Finance authority	Scenario calculation; optional narrative support	Feasible transfer; portfolio contribution and cash
Should we scale or stop?	Mature readout and full operating cost	Accountable commercial owner	Independent method review; AI may explain signed results	Replicate, revise or close; subsequent observed outcomes

Use the Rule / Model / AI / Agent / Human router. Rules handle exact criteria and repeatable calculations. Statistical models estimate risks or outcomes and must be validated. Generative AI interprets unstructured context and drafts inspectable work. An agent selects permitted next steps and tools dynamically; it needs a separate advantage over a fixed workflow. Humans retain ambiguous and consequential judgement. These are choices within a task, not a maturity ladder.

The operating records are deliberately small: T1 intervention and investment card; T2 economics and benefits ledger; T3 decision, experiment and evidence record; T4 action authority and operating contract; T5 delivery and retirement plan. Use existing systems or linked sections of one document. Do not create five new applications.

Your next move

Follow one recent recommendation to its actual action and outcome. If it stopped at a dashboard, identify whether authority, treatment, capacity or evidence was missing.

Evidence for this chapter3claims reviewed, with confidence ratings, limits and sources.

Claim classifications

Accepted means a principle or mechanism survived critical review, not that it guarantees a return. Accepted with conditions means prerequisites apply. Qualified means the original claim has been narrowed. Advanced needs substantial readiness. Experimental needs a capped comparison. Unresolved means the evidence cannot settle the claim. Rejected means the recommendation is excluded.

High, Medium and Low describe confidence in the stated target, not the chance of a positive ROI. For a Qualified or Rejected historical claim, confidence applies to the decision to narrow or reject it and to the supported limit—not to the original headline. Read the headline and classification together. An accounting identity or design principle is not empirical evidence of a commercial outcome. Follow the linked chapters, Plays and tools to see where each claim applies.