Startup Metrics: Build a Stage-Appropriate Evaluation Scorecard

Before launch, focus on the assumptions that determine whether the product should exist: a clear customer problem, feasibility and evidence from realistic tests. After launch, investigate whether customers reach useful outcomes and return. With repeatable paid use, examine the cost and reliability of acquiring and serving more customers. These are planning choices, not mandatory stages with universal performance thresholds.

Write the decision beside each metric. For example, “time to first completed task” can guide an onboarding change, while “cash available under the current spending plan” informs hiring timing. If nobody can name a decision that a number affects, it may belong in a diagnostic report rather than the main scorecard. Keep a few critical risks visible even when they cannot be reduced to a percentage.

For every measure, record its name, calculation, data source, reporting period, exclusions and owner. Define the customer unit: account, paying organization, site or individual user. Changing that unit between reports can create apparent growth without any underlying improvement. Show missing data explicitly and do not treat an unmeasured result as zero.

Stripe’s Billing Analytics documentation illustrates why definitions matter: settings such as discount treatment and active-subscriber timing can affect reported metrics. Align internal reports before comparing a finance spreadsheet with a dashboard. Keep recurring revenue separate from one-off fees and cash collection, and preserve the definition used for each reporting period.

For product behavior, identify the meaningful starting and return events. Amplitude’s retention guide makes those choices explicit. An account returning to finish a workflow is different evidence from an automated background event firing.

An original starter template is below. Select the rows that fit the business and replace generic measures with precise definitions. A transactional marketplace may care about successful repeat transactions, while a subscription product may need renewal cohorts. Avoid forcing both into an identical target. Targets should reflect the business plan and available evidence, with a date for reassessment.

Imagine a hypothetical subscription cohort with twenty paying accounts and $2,000 of monthly recurring revenue at the start. Three months later, four original accounts have left, removing $400 of that recurring revenue. The remaining accounts have added $600 through expansion. Assume no other changes, reactivations or currency effects, and exclude all newly acquired accounts from this cohort calculation.

Customer retention is sixteen divided by twenty, or 80%. Net revenue retention for the same original cohort is ($2,000 − $400 + $600) divided by $2,000, or 110%. Both numbers are correct under these definitions. Revenue expansion has more than offset losses financially, but it has not made the departed customers disappear.

The next action is to investigate who left and who expanded. If expansion comes from one large account, the company may have both customer-loss and concentration questions. If onboarding problems explain departures, the operational row should guide a test. Do not use the higher revenue percentage to declare every part of the business healthy.

Hold a regular review using comparable periods and mature cohorts. Check whether a change reflects customer behavior, a measurement change or an unusual event. A new definition can be useful, but document it and avoid silently rewriting historical results. When samples are small, show counts and customer context rather than presenting noisy changes as a reliable trend.

NextUnicorn’s published Rating framework describes using factor-level assessments to discuss strengths and areas needing attention. Keep an operating scorecard grounded in its underlying records even when using a broader assessment framework. A provider’s grade does not replace the company’s metric definitions or establish independent validation.

At the end of the meeting, select the issue with the highest decision value. Assign one investigation, product change or operational experiment, then record what evidence would make the team reconsider. This turns startup KPIs into a working management tool. For the deeper question of durable customer value, use the separate product-market-fit guide.

Explore NextUnicorn’s published Rating framework