Quick answer
A North Star metric is the repeatable customer behavior most closely linked to lasting product value. For early SaaS, it should be understandable, measurable, and difficult to improve through activity that does not help users.
Key takeaways
- A value metric needs an explicit success and deduplication rule.
- Pair activity with quality and retention safeguards.
- Treat the first metric as a hypothesis you can test.
Separate customer value from company revenue
Monthly recurring revenue matters commercially, but it is often a lagging outcome. A product that automates invoices might measure invoices successfully processed by active organizations. An analytics product might examine websites that generate usable reports and return to those reports.
Avoid choosing pageviews or registrations simply because those numbers are easy to grow. More account creation does not prove that users experienced value. The proposed metric should connect directly to the product promise.
Test candidates against four criteria
Ask whether the metric reflects a real customer outcome, can move within a useful timeframe, is consistently instrumented, and correlates with retention in your own data. Segment by customer type if sufficient evidence exists and compare retention for cohorts that reach the milestone versus those that do not.
Document the event definition and exclusions. For a “weekly activated team,” specify which actions qualify, how team identity is resolved, and whether automated system events count.
Pair the metric with guardrails
A team could increase “reports generated” by automatically creating reports nobody reads. Guardrails prevent this kind of false improvement. Useful companions include customer retention, time to first value, error rate, support load, and paid conversion.
If weekly active reporting teams grow 20% while trial-to-paid conversion falls and complaints double, the North Star is not telling the whole story. Review both movement and quality.
Revisit the choice as the product matures
At launch, you may have too few customers to validate a robust metric. Treat the first selection as a hypothesis, collect qualitative feedback, and revisit it after several cohorts. Do not change the definition every week; that destroys trend comparability.
Score a candidate before adopting it
For an illustrative invoice tool, compare two candidates: invoices created and invoices successfully delivered by active accounts. Creation can increase through retries, abandoned drafts or automation. Successful delivery is closer to the customer's task, but still needs a duplicate rule.
Write a small specification: count an account once per calendar week when at least one non-test invoice reaches the delivered state; exclude internal accounts and repeated notifications; retain a separate delivery-error measure. Record whether the business works at account or person level. A team of eight people should not become eight customer accounts.
Then inspect several completed cohorts in your own product records. Does achieving the milestone precede continued use, or are the most active customers simply more likely to do everything? The relationship is a hypothesis, not causal proof. Revfaro can collect a deliberately emitted goal; account-level cohort calculations may still belong in your product database or a worksheet.
Related reading
- Visitors vs Sessions vs Pageviews vs Events: A Practical Analytics Guide
- The Essential SaaS Analytics Dashboard: 12 Metrics Worth Reviewing
- Website Analytics for SaaS Founders: What to Measure First
Frequently asked questions
Is MRR a good North Star metric?
It is a crucial business KPI but usually a lagging outcome. Pair it with a customer-value metric that guides product decisions sooner.
Should every feature have its own North Star?
No. Teams can use supporting product metrics, but an overall product benefits from one clear value measure with transparent guardrails.