ICP & Positioning

What this is for

Give the whole team one specific, reviewable answer to who the product is for, what progress that customer wants, and which evidence supports the belief.

When this helps

• ICP documents become polished guesses with no source trail.
• Product, sales, and marketing use different definitions of the customer.
• Broad personas do not tell a PM which opportunity, experiment, or message to choose.
• New evidence arrives but the customer definition never changes.

How it works

ICP & Positioning stores one working profile per product workspace in the Product Graph. You can write the profile manually or ask Specky to draft it from real, non-example signals. The profile covers target companies, buying roles, jobs, pains, trigger events, desired outcomes, objections, anti-ICP, customer language, and distinct segments. Every segment carries a fit score and a separate evidence-strength label, because strategic attractiveness and confidence are not the same thing. Generated citation IDs are checked against real workspace nodes before the profile is saved.

Before you start

• Your current belief about the best-fit company, user, buyer, problem, and outcome.
• At least one real signal when possible: an interview, feedback item, call, support conversation, CRM note, analytics observation, or product artifact.
• The active product workspace. ICPs are product-specific so multi-product organisations do not mix audiences.
• Human judgment about which segment deserves focus and which claims remain assumptions.

What you should see

• An editable ICP snapshot and profile-readiness score.
• Primary, secondary, and experimental segments with fit and evidence strength.
• Jobs, pains, triggers, outcomes, objections, customer language, and anti-ICP.
• A visible evidence inventory by source with recent signals Specky can cite.
• Direct actions into customer research, experiments, Planning Studio, and Feature Marketing.

Try it now

1. Open Market → ICP & positioning. This view is always visible, including in existing workspaces with older sidebar preferences.
2. Write one sentence describing who you believe gets the most value and why. Specific buying situations are more useful than broad demographics.
3. Choose Build ICP from evidence. Specky reads available real signals, drafts the profile, and marks it as learning rather than validated.
4. Open Evidence and inspect the sources. Check that the customer language and segment distinctions are supported; anything unsupported should stay an assumption.
5. Choose Edit ICP to correct the profile, add missing industries, stages, jobs, pains, triggers, outcomes, or anti-ICP, and set the confidence state.
6. Open Activate and carry the selected customer definition into research, an experiment, planning, or a feature launch.
7. Refresh from evidence when meaningful new interviews, calls, feedback, or behaviour arrive.

A real example

A founder starts with “B2B SaaS product teams.” Interviews and support evidence show the strongest recurring situation is a growth PM at a 20–200 person SaaS company investigating an activation change across PostHog, Slack, and customer conversations. Specky creates that as the primary segment, keeps enterprise leadership as a secondary segment, and flags buying authority as an evidence gap instead of inventing certainty.

Your next move

Use Feature Marketing to turn a segment and product change into a customer-facing launch, or Alex Research to fill a weak evidence area.