PMF Dashboard Template

Make the Week End in One Decision

A pre-PMF dashboard should make it difficult to tell a comforting story. If signups rise while qualified accounts fail to reach value, the failure should be visible. If customers return and pay only because the founder repairs every job, that labor should remain attached to the result. If one segment is working and another is not, an average should not blend them into ambiguity.

Use this template once the metrics tree has named the target segment, primary value behavior, and natural frequency of the customer’s problem. Use the event tracking plan to establish which product facts can be trusted. The dashboard does a different job: it brings the evidence together at a fixed review time and forces one operating decision.

Do not begin in charting software. The first version can be a text document or spreadsheet backed by an account ledger. It is complete when it helps you choose what to build, repair, investigate, narrow, price, or stop—not when every available event has a panel.

Write a Review Contract

At the top of the dashboard, write the claim you intend to judge. Include one target segment, the behavior that represents delivered value, the natural interval at which that value should recur, and the uncertainty currently blocking a decision.

For a product that turns advertising exports into client-ready weekly digests, the contract might be:

Every Monday, judge whether small agencies can send a client-ready digest each week without export cleanup by the founder. Compare founder-recruited agencies with public-launch signups. This month, decide whether the import path is ready for more acquisition or still needs to become self-serve.

That statement prevents a favorable number from silently changing the question. Total reports generated cannot substitute for client delivery. Activity from every signup cannot dilute the chosen segment. An assisted success cannot quietly become a product-delivered success.

Keep the contract stable for an observation window. Change the active uncertainty as evidence moves; change the segment or value behavior only when you are deliberately changing the product thesis.

Keep an Inspectable Account Ledger

The dashboard is the summary. Underneath it, keep one history for each account or other value-receiving unit. A user is not always the right unit: a clinic, workspace, project, repository, or agency may receive the outcome even when several people touch the product.

For each unit, preserve enough history to reconstruct the weekly reading:

• segment, use case, acquisition source, and qualification state;
• first value behavior, time-to-value, and any founder assistance;
• the next eligible opportunity for value to recur and what happened;
• payment, discount, renewal, cancellation, and custom commitments;
• support time, manual work, failures, and operational interruptions;
• customer language, referrals, expansion, and churn tied to behavior;
• product or experiment version encountered by the account.

The ledger may combine product data, payment records, support notes, and manual observation. Mark those boundaries. Manual evidence is useful; unmarked manual intervention is misleading. Preserve old definitions and segment labels when they change so that a repaired record does not masquerade as product improvement.

Assemble the Weekly Page

Copy the following page. Replace prompts with current evidence, ordinary language, and inspectable account identifiers. Raw counts are often more honest than percentages when the sample is small.

PMF DASHBOARD — [TARGET SEGMENT] — [REVIEW DATE]

REVIEW CONTRACT
Primary value behavior:
Natural frequency and next eligible opportunity:
Current uncertainty:
Comparison that could reveal a hidden difference:

ACTIVATION — did the right accounts reach first value?
Qualified starting accounts:
Accounts reaching first value / window:
Time-to-value:
Assisted versus unassisted:
Where qualified accounts stopped:

REPEATED VALUE — did value recur on the customer's clock?
Accounts eligible to repeat:
Accounts repeating the value behavior:
Accounts requiring a reminder or founder rescue:
Strongest retained and non-retained account to inspect:

REVENUE QUALITY — did money follow repeatable value?
Payment, renewal, or expansion attached to repeated value:
Discounts, custom promises, refunds, or churn:
Cost or delivery work that could reverse the reading:

QUALITATIVE PULL — what does behavior-backed language explain?
Deadline, dependency, complaint, invitation, referral, or expansion:
Account and observed behavior attached to that language:
Current explanation supported or challenged:

FOUNDER LOAD — can one person keep the promise?
Support and manual minutes by cause:
Failed jobs, escalations, and interruptions:
Accounts whose success depended on founder work:

ACTIVE EXPERIMENT
Change and exposed accounts:
Expected movement:
Observed movement:

EVIDENCE CONFIDENCE
Missing, late, duplicated, assisted, or redefined evidence:
Accounts or source records inspected:
Claim the evidence cannot yet support:

ONE DECISION

EXPECTED MOVEMENT BY NEXT REVIEW

ACCOUNT OR QUESTION TO INVESTIGATE FIRST

Each section carries a distinct question. Activation asks whether qualified accounts reach value at all. Repeated value waits for the customer’s next real opportunity rather than using a convenient calendar window. Revenue quality connects money to repeatable delivery. Qualitative pull helps explain behavior instead of replacing it with praise. Founder load exposes the work the product has not absorbed.

The comparison line should test a live explanation: segment, acquisition source, onboarding path, product version, or assisted versus unassisted use. Do not add a breakdown merely because the data permits it.

Read One Week Without Averaging Away the Problem

Suppose twelve small agencies entered the reporting workflow. Six sent a first client digest; three of those needed export cleanup by the founder. At the next weekly opportunity, four sent another digest, but one again needed a repair. Three retained agencies continued paid pilots. Support consumed 164 minutes, 118 of them on import cleanup.

Several readings are true. Half the agencies activated. Four of six activated accounts repeated the value behavior. Payment followed repeated use. Yet most support time accumulated before the promise could be delivered without help.

The dashboard should not resolve that tension with a blended PMF score. Open the account histories. If founder-recruited agencies arrived with cleaner exports while public-launch accounts brought unsupported formats, the next decision could concern qualification or the product promise. If qualified accounts with the intended format still required repair, import validation is the more plausible constraint.

A useful weekly decision would be:

Pause broad acquisition and make the supported import path self-serve. In the next four qualified agencies, expect a first digest to be sent without data cleanup by the founder.

This decision names authorized work, a boundary, and observable movement. “Add an import dashboard,” “improve activation,” and “watch retention” do not.

Run the Review in a Fixed Order

Review the dashboard at the cadence of the customer problem, at the same time each cycle. Begin with evidence confidence. A missing core event, mixed account and user counts, an unmarked assisted cohort, or a changed definition may make the week’s first task a repair to the record.

Then read activation, repeated value, revenue quality, qualitative pull, and founder load. Find the weakest link in the current value path. Inspect the accounts behind it, including the account that most strongly contradicts your preferred explanation.

Before choosing work, write two or three plausible causes. Distinguish them with the cheapest evidence already available: a failed trace, support message, payment record, or account conversation may settle the question more directly than another metric. End the review with one decision and the movement you expect by the next eligible cycle.

Let the Spine Change Slowly

Keep the review contract, account ledger, five evidence sections, definition history, and decision log stable enough to compare. Experiments and active uncertainties can change more often.

Add a measure when the same decision has been blocked twice because the evidence was absent. Move a diagnostic out of the weekly view when its question has been answered. Retire a measure when it no longer changes action. When a definition changes, date the change and do not draw an uninterrupted trend across it.

Be suspicious of these dashboard failures:

• a total without the target segment or eligible denominator;
• activation defined as setup rather than the first honest value behavior;
• retention measured before the customer’s problem could recur;
• payment detached from discounts, custom work, or delivery cost;
• praise detached from what the customer did;
• founder rescues counted as unassisted product success;
• a composite PMF score that conceals the evidence underneath;
• a review that ends with several priorities or no decision.

The dashboard is ready when every reading can be traced to accounts, founder work remains visible, uncertainty is stated without false precision, and the page ends with one bounded choice. Its purpose is not to certify product-market fit. Its purpose is to reveal the next constraint before the founder spends another cycle avoiding it.