Product-led growth produces an enormous amount of data and very little insight by default. Every click can be logged, so every click gets logged, and the dashboard fills with numbers that move without meaning anything.
A workable PLG measurement system is small. Five metrics, each tied to a decision someone has to make.
1. Activation rate
The share of new signups who complete the action that predicts retention. Not “completed onboarding” — a specific outcome in the product, like publishing the first page, connecting the first data source or inviting the first teammate.
This is the single most important PLG number because everything downstream depends on it. A signup that never activates will not convert, will not expand and will not refer. Doubling activation does more for revenue than doubling traffic, and it is usually cheaper.
Finding the right action takes analysis rather than opinion. Compare users retained at day 90 against users who churned in the first month, and look for an early action where the two groups diverge sharply.
2. Time to value
How long from signup to that activation moment. Measured in minutes for simple products, days for complex ones.
Activation rate tells you whether people get there; time to value tells you how hard it is. A product with 60% activation at a median of four hours has a different problem from one with 60% activation at a median of nine days. The second is losing people to competing priorities rather than to confusion.
Track the median, not the mean, and plot the distribution. A bimodal distribution usually means you have two user types with very different needs.
3. Product-qualified leads
A PQL is a user whose in-product behaviour indicates readiness to buy or expand: hitting a plan limit, inviting several colleagues, using a premium feature during a trial, reaching a usage threshold.
The definition is the hard part, and most companies make it too broad. A PQL definition that flags 30% of your free users is not qualification, it is a mailing list. Start narrow — a definition that fires for 3–5% of users and converts at a high rate — and widen only once sales can keep up.
The test of a good PQL definition is simple: does a salesperson, shown the account, agree it is worth a call? If not, tighten it.
4. Natural rate of growth
What share of new annual recurring revenue arrives without sales involvement. Self-serve signups, organic expansion, referrals — revenue the product acquired on its own.
This is the metric that tells you whether you are genuinely product-led or just have a free trial. A company where 15% of new ARR is self-serve is sales-led with a trial on the website. A company at 60% has a product that sells.
It also matters for planning: self-serve revenue scales with product and traffic investment, sales-led revenue scales with headcount. Knowing the mix tells you which budget line creates growth.
5. Expansion by cohort
How much more a cohort of accounts pays at month twelve than at month one. This is cohort-level net revenue retention, and in PLG it is where most of the economics live, because initial contract values are small by design.
Track it by cohort rather than in aggregate so you can see whether product changes are improving it. Aggregate NRR moves for too many reasons at once.
The metrics to stop reporting
Four numbers appear on nearly every PLG dashboard and earn their place on almost none.
Total signups. Goes up when you spend more on ads, says nothing about whether the product works. It is an input, not an outcome, and treating it as a goal produces low-quality signups that depress every other metric.
Daily active users as a headline. Useful for consumer products, misleading for B2B tools used weekly by design. A project management tool should not expect daily use from everyone, and chasing it leads to notification spam.
Feature adoption percentages. Interesting to the team that built the feature, rarely decision-relevant. The useful version is whether adopting the feature changes retention, which is a different question.
NPS as a primary metric. Slow-moving, sample-biased and uncorrelated with revenue in most B2B contexts. Worth collecting for the free-text comments, not worth putting on the wall.
Building the funnel view
The five metrics assemble into one chart: visitors to signups to activated to retained at day 30 to converted to expanded. Each stage with a conversion rate between it and the next.
Reviewed monthly, this view answers the only question that matters in a planning meeting — which stage is leaking most — and it prevents the common failure of optimising the stage that happens to have an owner rather than the stage that is broken.
Segment it by acquisition source once you have the volume. Organic signups, paid signups and referrals behave differently enough that a blended funnel can hide a channel that is genuinely not working.
Instrumenting it without a data team
Three practical notes for teams doing this with limited engineering support.
- Define events before you implement them. A written list of twenty events with precise definitions, agreed by product and marketing, prevents the usual outcome of 400 events nobody trusts.
- Put the user and account identifier on everything. B2B analysis is almost always account-level, and retrofitting account identity onto user-level events is painful.
- Store raw events somewhere you own. Analytics tools change; your event history should not be locked inside one. Even a simple warehouse table is enough.
When PLG metrics mislead
Two situations where the standard PLG dashboard gives the wrong answer.
The first is a product sold to teams but adopted by individuals. Individual activation looks excellent, revenue does not follow, because the person who activated cannot buy. The metric to add is account-level activation: has anyone with budget authority engaged?
The second is a product with a long natural usage cycle — annual planning tools, tax software, hiring systems. Day-30 retention is meaningless when the product is used in bursts. Measure retention against the natural cycle instead, even if that means waiting a year for clean data.
Connecting PLG metrics to revenue
The gap between a PLG dashboard and a revenue forecast is usually a missing translation layer. Three ratios bridge it.
Signups per thousand visitors. Ties marketing spend to pipeline. If this moves without traffic moving, something changed on the signup page or in the offer.
Paying customers per hundred signups. The compound of activation, retention and conversion. This is the number to forecast from, because it already absorbs the funnel’s internal losses.
Average revenue per paying customer at month twelve. Not at month one. PLG accounts land small and grow, so a forecast built on initial contract value will understate revenue by a wide margin.
With those three, a traffic number becomes a revenue number, and a product improvement can be argued for in the same currency as an advertising budget. That is usually the missing piece when a product team struggles to get investment for onboarding work.
A sensible review cadence
Weekly, look at signups, activation rate and PQL count — the fast-moving operational numbers. Monthly, look at the full funnel with conversion rates between stages, segmented by channel. Quarterly, look at cohort expansion and natural rate of growth, which only make sense over longer windows.
Mixing these cadences is the common error. Reviewing cohort expansion weekly produces noise and anxiety; reviewing activation quarterly means a broken onboarding flow runs for three months before anyone notices.
Frequently asked questions
How many metrics should a PLG dashboard have?
Five to seven on the main view. Anything more and nobody reads it; anything fewer and you cannot diagnose where a problem sits.
What is a good activation rate?
It depends entirely on how the milestone is defined, which makes cross-company comparison meaningless. Judge it against your own trend and against the gap between activated and non-activated retention.
Is PLG incompatible with a sales team?
No. Most successful PLG companies add sales for larger accounts, with PQLs as the handoff. The product does the qualification and the first demonstration of value; sales handles the commercial conversation.
Should free users count in retention metrics?
Report them separately. Mixing free and paid retention makes both numbers uninterpretable, since the two groups behave and churn for different reasons.
