Lean Analytics Stages
Five stages every startup must navigate from idea to scale
The Lean Analytics Stages framework identifies five sequential phases that startups pass through: Empathy, Stickiness, Virality, Revenue, and Scale. Each stage has distinct goals, metrics to track, and gates that must be cleared before advancing. The framework synthesizes insights from Dave McClure's Pirate Metrics, Eric Ries's Engines of Growth, Sean Ellis's Startup Growth Pyramid, and Ash Maurya's Lean Canvas into a unified progression model.
The critical insight is that premature advancement kills startups. Investing in viral growth before your product is sticky wastes money on users who immediately leave. Scaling revenue before you have viral mechanics means paying full price for every customer. Each stage builds on the foundation of the previous one, and skipping ahead is the startup equivalent of building on sand.
At each stage, you pick the One Metric That Matters and optimize it until you clear the gate. In Empathy, you validate that a painful problem exists. In Stickiness, you prove users keep coming back. In Virality, you get existing users to bring in new ones. In Revenue, you prove the unit economics work. In Scale, you prove the business can grow independently across markets and channels.
- Each stage must be completed before advancing to the next; premature advancement is the most common startup killer
- The metrics that matter change fundamentally at each stage, from qualitative empathy signals to quantitative scale economics
- Gates between stages are defined by specific metric thresholds that prove readiness to advance
- Your business model and your current stage together determine which specific metric to track right now
- You never fully leave a stage; earlier metrics remain important as guardrails even as your focus shifts forward
- Empathy: Validate the problemConduct problem interviews with at least 15 prospective customers. Discover whether a genuine, painful problem exists and whether your proposed solution resonates. Your metrics here are qualitative: interview scores, patterns in feedback, and willingness of subjects to refer others or pay immediately.Pro tipScore your problem interviews on a consistent rubric covering problem ranking, existing solution attempts, engagement level, willingness to do a follow-up, referrals offered, and offers to pay.WarningDo not skip qualitative research because it feels unscientific. Quantitative data cannot tell you why something is happening.
- Stickiness: Prove users keep coming backBuild your MVP and focus squarely on retention and engagement. Track daily and monthly active users, time between visits, feature usage, and churn. Use cohort analysis to see if your iterations are improving stickiness over time. Do not invest in growth until you can prove that users who arrive actually stay.Pro tipIf you cannot convince 100 users to stick around today, you are unlikely to convince a million later. Start with a tiny, hyper-focused target market.WarningDo not drive new traffic until you know you can turn attention into engagement. Premature growth burns money and time.
- Virality: Get users to bring in other usersFocus on your viral coefficient, invitation rates, acceptance rates, and viral cycle time. Distinguish between inherent virality built into product use, artificial virality from incentives, and organic word-of-mouth. Optimize the viral loop by measuring each step and removing friction.Pro tipViral cycle time may matter more than viral coefficient. Halving the time it takes users to invite others can be the difference between 20,000 users and 20 million.WarningWatch that virality does not come at the expense of engagement. New users acquired virally may differ from early adopters and may not stick.
- Revenue: Prove the unit economics workShift focus to customer lifetime value, customer acquisition cost, and the ratio between them. Aim for CLV to be at least three times CAC. Track time to customer breakeven and monthly recurring revenue. Experiment with pricing, tiers, and upselling to maximize revenue per customer.Pro tipCustomer acquisition payback in months is a powerful single number that rolls up marketing efficiency, customer revenue, cash flow, and churn rate.WarningDo not just look at top-line revenue. A customer that costs more to acquire than they generate in lifetime value will bankrupt you no matter how fast you grow.
- Scale: Prove the business can grow across marketsCompare metrics like customer acquisition payback across channels, regions, and campaigns. Look at ecosystem health, channel partner effectiveness, and competitive positioning. Use the Three-Threes Model to maintain discipline: three big assumptions at the board level, three tactical actions at the executive level, and three experiments at the execution level each week.Pro tipUse Porter's generic strategies to decide whether you are competing on efficiency or differentiation. Trying to do both simultaneously at this stage is dangerous.
Buffer started by validating demand, then focused on stickiness metrics: 60 percent of signups returning in the first month, 20 percent still active after six months, and churn hovering around 2 percent. Only after proving stickiness and reaching ramen profitability did founder Joel Gascoigne shift the OMTM from revenue to user acquisition.
Backupify was paying 243 dollars to acquire a customer who paid only 39 dollars per year. The unit economics were fundamentally broken for consumer sales because backup services are not naturally viral. The company pivoted from consumer to enterprise sales.
Croll and Yoskovitz developed this framework after reviewing and comparing multiple existing startup frameworks, including Pirate Metrics (AARRR), Eric Ries's three Engines of Growth, Sean Ellis's Startup Growth Pyramid, and the Long Funnel concept. They found that while each framework offered valuable perspectives, none provided a clear stage-by-stage progression with explicit gates that told founders when to advance.
Their experience at Year One Labs, working hands-on with multiple startups at similar stages, revealed a consistent pattern: the most common failure mode was premature advancement, particularly spending money on growth before achieving product stickiness. This led them to create a model that explicitly sequences the stages and defines measurable gates between them.