Marketing

Which marketing decision would create the clearest evidence of useful demand?

Marketing connects a valuable offer to people or agents who can benefit from it. A complete marketing capability needs five views: Principles constrain the work; Processes define its jobs; Platform enables them; Players own judgment; Performance tests whether the work created value.

Choose the missing view

Question you cannot answer	View to use	Output
What must never be traded away for reach?	Principles	A decision rule and kill signal.
Which repeatable marketing job owns this outcome?	Processes	One job, trigger, inputs, method, and receipt.
What data, software, protocol, or integration enables the job?	Platform	A capability requirement before a product choice.
Who decides, assists, approves, and remains accountable?	Players	A human/AI responsibility boundary.
What observation will show progress, failure, or harm?	Performance	A baseline, gauge, review date, and stop condition.

Use all five views to specify the capability. Use only the missing view to repair the next decision.

Choose one of eight jobs

Situation	Canonical job	Required output
The audience is too broad or assumed.	Define an ideal customer profile	A bounded customer, situation, need, alternatives, and disqualifiers.
Buyers or agents cannot reliably discover and classify the offer.	Optimise agent intent	A sourced question map and discovery-to-action receipt.
Publishing is sporadic or purely promotional.	Execute a content calendar	A useful sequence with owners, dates, and evidence to review.
Technical depth is valuable but hard to understand.	Position protocol depth	A truthful value translation with proof and limitations.
Machines can discover the offer but cannot evaluate or act.	Enable agent-readable conversion	A machine-readable path from claim to evidence and authorised action.
Repeat participation should compound shared value.	Design onchain loyalty	An incentive model with consent, abuse tests, and a stop condition.
Partners or communities can carry value further.	Distribute through ecosystems	A bounded partner route with mutual value and ownership.
A consented onchain outcome must be connected to a source.	Attribute onchain outcomes	An attribution method with privacy and confidence bounds.

These are stable job identities. Supporting tactics do not become new jobs.

Compose a go-to-market cycle

When the individual jobs are clear but the route is fragmented, use GTM Distribution. It composes the jobs into one cycle:

frame one audience and value bet
→ choose one channel wedge
→ make discovery and conversion possible
→ deliver a useful sequence
→ respond with human ownership
→ compare outcomes with the bet
→ recommit, change one variable, pause, or stop

Start with the public GTM Strategy instrument when the offer, market, positioning, or route is still unsettled.

Human and AI responsibility

Humans own the beneficiary, promise, public claim, trade-off, spend, consent, consequential response, and stop decision. AI may search and organise evidence, compare alternatives, draft variants, inspect machine readability, classify signals, and prepare a review.

Useful AI capability depends on the job:

• use retrieval and citation support when claims depend on changing evidence;
• use structured extraction and classification for repeated signal review;
• use generation for drafts and variants, never as proof;
• use evaluation to compare outputs against explicit criteria;
• use tools only through authorised data, publishing, spending, and contact boundaries.

Do not automate a marketing task whose beneficiary, evidence standard, privacy boundary, or accountable human is unknown.

Proof and failure

Begin with a baseline and one beneficiary outcome. Route signals such as reach, visits, citations, replies, wallet connections, and sign-ups can explain how value travelled. They do not prove that the beneficiary improved.

The cycle succeeds when the intended audience can understand the offer, take a useful next action, and produce evidence that improves the next decision. It fails when activity grows but the team cannot connect it to useful demand, learning, or a justified stop.

Stop or reframe when:

• no specific audience, situation, or useful outcome is named;
• the output relies on unsupported or synthetic customer evidence;
• channels multiply before one route returns evidence;
• tracking lacks consent or confidence bounds;
• AI can publish, spend, contact, or commit without accountable review;
• the review rewards activity while beneficiary outcomes remain unknown.

Do it now

Choose one row from the eight-job table. Write its required output in one page. Add the evidence you have, alternatives considered, accountable human, one gauge, one stop condition, and a review date. Do not add a tool until the page shows what capability the tool must enable.

At review, keep observations separate from interpretation. Record what changed your next decision. Then continue independently through the relevant process page, use its linked public instrument, or request a bounded marketing review if the remaining judgment carries material consequence.

Context

• depends-on Business — choose the rightful decision route before selecting tactics.
• depends-on Distribution — carry value to a beneficiary rather than optimise attention alone.
• applies-to Marketing Processes — use the eight canonical jobs and their supporting methods.
• applies-to GTM Distribution — compose the required jobs into one closed route.
• pairs-with Marketing Instruments — make the decision concrete and reviewable.
• proved-by Marketing Performance — distinguish route signals from beneficiary outcomes.

Questions

Next question: Which job can return the most decision-changing evidence in one cycle?

• Which of the five views rests on assertion rather than evidence?
• What judgment must remain human-owned?
• Which signal would make you stop rather than add another channel?