FDI method

Growth Cartography is the method behind Marketing Engineering: evidence first, one supervised loop, measurable handoff.

Signal Audit / Owned-demand loop

Start with the problem that already has a buyer.

FDI is testing one owned acquisition loop with one buyer segment: B2B marketing leaders at $10M–$500M companies who have an active demand, AI, or measurement program but no trusted path from signal to qualified pipeline.

Signal Audit

What the answer engines say about you, and whether it is true

A rank is the wrong number. An answer that names you first for the wrong reason loses the deal as fast as one that never names you at all.

Built for B2B marketing leaders carrying a revenue number, roughly $10M to $500M.

Measured on ChatGPT, Claude, Perplexity and Google AI Mode.

  • Mentioned

    Are you in the answer at all, across a fixed panel of the questions your buyers actually ask. Measured per platform, so a gap on one engine is visible instead of averaged away.

  • Accurate

    When you are named, is what it says true. A region limit, a stale claim or a category confusion filters you out before a human ever sees you.

  • Influence

    Does the mention change the recommendation, meaning the mention changes what they pick. This is the only version of visibility that shows up in pipeline.

Each answer is checked against claims we can verify. A region limit, a stale claim, or a category confusion is recorded as a flag.

The visits your analytics counts as direct

Answer engine visits arrive with no referrer, so they are counted as direct traffic. We isolate them behaviourally: direct traffic to long tail pages that now behaves like search traffic, measured against your own pre AI baseline.

This is an estimate from observed behaviour, not a referrer measurement. We report a range because the honest answer is a range.

Answer freshness

A change on your site does not reach the answer on the same day. Each platform caches for a different period, and it varies by content. We re measure on that clock rather than on a calendar, and report how long your change took to appear.

Illustrative panel, not a client result: named in 3 of 12 prompts, one accuracy flag, recommended in none.

What you receive

  • The prompt panel, versioned, so two audits are comparable
  • The raw answers, archived with capture dates and cited sources
  • The scored panel: mention rate, accuracy flags, influence grade
  • The dark traffic estimate, with the method and the limits stated
  • A re measurement date, keyed to each platform's cache window

One analytics export, a 30 minute session to agree the prompt panel, and one review pass.

We do not sell citations and we do not promise rank. If the panel says you are absent, the audit says you are absent.

Request the audit

Five bottom-funnel entry paths

Match the problem to the next decision.

Each page names one high-intent problem, one useful action, one inquiry source, and one human follow-up path. Attention is diagnostic. Qualified inquiry is the success measure.

content to pipeline

Owned content earns attention but not qualified demand

You can point to traffic, rankings, or engagement, but not to a trusted sequence from a high-intent problem to a human-reviewed inquiry.

marketing measurement

Marketing activity is visible, but the qualified path is not

The team debates channel performance because clicks, form records, qualification, and pipeline live in separate stories.

AI search visibility

AI search visibility is increasing, but authority is not traceable

The team sees impressions, mentions, or citations but cannot show which owned evidence was retrieved, what the buyer did next, or where a human should intervene.

AI marketing governance

AI-assisted marketing output is growing faster than review and ownership

The team has more drafts or automations, but unclear source boundaries, approval ownership, correction paths, or evidence that output improves a business outcome.

Address

Advertising is changing address, not shrinking

The budget is still planned by channel, but nobody can say which share lands where a human is present, which surfaces an agent could substitute, or whether your own properties are legible to a machine.

What happens after inquiry

Evidence first. Qualification by a human.

The form records the source, campaign, landing page, referrer, and influenced page with the inquiry. FDI then reviews fit, need, authority or access, and a plausible next step before recording a qualified inquiry.

01 · Capture

One meaningful event

Each page submits the canonical inquiry_submitted event. A click or page view never counts as qualification.

02 · Follow up

One useful reply

FDI reads the context, asks the shortest set of fit questions, and keeps the thread human.

03 · Review

One weekly decision

Review inquiry quality by source and page, then change one hypothesis before adding more distribution.

Signal Audit + Agent-Boundary Governance

Governance Package

Two problems arrive together right now: marketing systems produce signals nobody can act on, and AI agents are starting to read the surfaces companies publish to. Fixing one without the other leaves the other half exposed.

Records-first diagnosis

The Signal Audit

The fixed-scope diagnosis of the evidence, conversion path, measurement gap, and smallest supervised system to build next.

Fail-closed protocol

Agent-Boundary Governance

A tested fail-closed protocol for what systems may safely expose to an AI agent: safe-only workbook architecture, minimum-disclosure rules, a preflight that blocks rather than warns, adversarial boundary testing, and a repeatable re-run protocol for platform behavior changes.

$4,500 fixed. The $2,000 audit-fee credit toward the four-week Marketing Engineering Sprint still applies if the Sprint begins within 30 days.