Lean Analytics Stages

---

Origins

Alistair Croll and Benjamin Yoskovitz introduced the Lean Analytics framework in their 2013 book of the same name.¹ The book's central argument: the right metric depends on where your product is. Early-stage and growth-stage products have entirely different success conditions, and tracking the same metric across both stages produces misleading signals.

Lean Analytics builds on the Lean Startup tradition of build-measure-learn cycles and extends it with a structured map of which metrics matter at which stage.

The Five Stages

Empathy

Earliest stage. The product doesn't exist yet, or exists only as prototypes. The team is trying to understand a real problem deeply enough to be worth solving. The "metrics" here are qualitative: interview count, problem-confirmation rate, willingness-to-pay signals.

The danger: skipping this stage and building before understanding. The cure: structured customer interviews until specific patterns emerge.

Stickiness

The product exists; the team is finding out whether anyone actually wants it. The metric that matters: do users come back? Retention curves, repeat-use frequency, and time-to-first-value dominate here. Acquisition is premature; revenue is dangerous.

If users don't come back, no amount of marketing fixes the product. The team's job is to make the product sticky before scaling anything else.

Virality

Sticky users are now bringing other users. The team is measuring the K-factor (viral coefficient) and looking for organic growth signals. Referrals, invites sent per user, share-to-conversion rate.

Not every product has a viral mechanic, and forcing one rarely works. For products that do, this stage is when the network effects start. For products that don't, the team moves to revenue.

Revenue

The product has users who stick and grow. Now the team optimizes the economics: customer lifetime value, acquisition cost, payback period. The classic LTV/CAC ratio belongs here.

Premature attention to revenue (before stickiness) corrupts the product. Late attention (well after stickiness) leaves money on the table.

Scale

Revenue economics work. The team's question is now: how big can this get? Market expansion, vertical or geographic extension, channel optimization. The metrics shift again to total-addressable-market signals.

The "One Metric That Matters" Discipline

The framework's hardest-hitting idea: at any given moment, the team should have one metric that matters more than all the others. Not three KPIs, not a balanced scorecard — one number, chosen because the team's most important learning depends on it.

Examples:

• Stickiness stage: D7 retention rate.
• Virality stage: viral coefficient K.
• Revenue stage: cohort LTV.

The single-metric focus is the same insight behind North Star Metrics, but Lean Analytics is more explicit about the metric changing with stage. The North Star is stable; the OMTM (one metric that matters) is stage-specific.

Why Stage-Matching Matters

The most common metric error is using the right metric at the wrong stage:

• Acquisition-stage metrics at stickiness stage: the team pours money into marketing while users churn after one visit. The growth looks good; the product is hollow.
• Revenue metrics at empathy stage: the team builds for what people will pay for before understanding what they need. Product-market fit is sacrificed for transaction count.
• Stickiness metrics at scale stage: the team obsesses over D30 retention while the market window closes around a faster competitor.

Common Failure Modes

• Choosing the wrong stage. Teams often claim to be in a later stage than they are. The honest test: is the metric of the previous stage actually solved?
• Skipping stages. Going from empathy directly to revenue (because that's what investors ask about) without confirming stickiness.
• Optimizing the wrong stage for too long. Continuing to push virality after the product has matured into a revenue-stage business.
• Treating "one metric that matters" as permanent. The metric is supposed to change as the stage changes. Teams that lock into one OMTM forever lose the framework's adaptive value.

Coaching Tips

Diagnose stage honestly.

Most teams claim to be a stage ahead of where they are. Test by asking whether the previous stage's metric is actually solved.

Pick one OMTM per quarter.

One metric. Whatever it is, the team obsesses about moving it. Multiple metrics produce hedged effort.

Don't skip empathy.

Teams that build before understanding the problem usually build the wrong thing. The empathy stage is the cheapest stage to invest in.

Resist premature revenue.

Optimizing revenue before stickiness corrupts the product. Wait until users come back before asking them to pay.

Switch the OMTM when you exit a stage.

The metric is stage-specific. When stickiness is solved, the team's attention moves — and the OMTM moves with it.

Use it for investor conversations too.

Naming your stage and your OMTM makes investor pitches sharper. "We're in stickiness; D7 retention is our metric" is more honest than a flat dashboard.

Summary

Lean Analytics is the most useful framework available for matching metrics to product stage. The discipline of choosing a single One Metric That Matters — and changing it as the product matures — keeps teams focused on the question they actually need to answer right now. Teams that skip stages or chase the wrong stage's metric burn time and capital. Teams that follow the framework rigorously discover which question they should be answering and answer it well before moving on.

Back to Agile Metrics