Most product-led growth teams have a growth problem they can't identify because they're not running a proper PLG funnel bottleneck analysis. They look at top-line signups and bottom-line revenue. They miss everything in between, which is almost always where the actual leak is.
This post walks through how to do bottleneck analysis the right way: measuring each stage, finding the constraint, and fixing it in an order that actually compounds.
A properly defined PLG funnel has four gates, not two.
Acquisition is when a new user creates an account. This is the metric most teams obsess over because it's tied directly to marketing spend. More budget in, more signups out. Simple to measure, simple to optimize.
Activation is when a user reaches the "aha moment" — the first time the product delivers its core value. Exactly what this means varies by product. For a project management tool it might be creating and assigning a task. For an analytics platform it might be running a first report. The key is defining it as a specific, observable user action, not a fuzzy feeling.
Engagement is when a user builds a habit around the product. They return multiple times within a defined window. They complete the core workflow repeatedly. Engagement is what separates users who stay from users who churn after one session.
Paid conversion is when a user moves from free to paid. This is the monetization event, and it's heavily influenced by everything that came before it.
Each transition between stages has a conversion rate. Acquisition to activation, activation to engagement, engagement to paid. Your bottleneck is the transition with the lowest conversion rate.
Start by pulling data on active users at each stage over a defined period, typically a 30-day cohort. You want four numbers: how many users entered activation, how many completed it, how many reached engagement, how many converted to paid.
The right tool for this is a product analytics platform like Amplitude or Mixpanel. Both let you build funnel visualizations that show drop-off at each step, segment by user properties, and set custom event definitions for each stage. If you haven't defined activation events in your analytics stack yet, that's your starting point before anything else.
Once you have conversion rates at each stage, lay them out in sequence. A typical healthy PLG funnel might look like 40-50% acquisition to activation, 60-70% activation to engagement, and 20-30% engagement to paid conversion. Your specific numbers will vary by category and ICP, but the shape matters more than the absolute numbers.
The bottleneck is the stage with the steepest relative drop. Not the lowest absolute number — the lowest conversion rate.
Here's an example: say 1,000 users sign up in a month. 550 activate (55%). 420 reach engagement (76% of those who activated). 60 convert to paid (14% of engaged users). The paid conversion rate of 14% is the floor, and that's where you focus.
But don't stop at identifying the stage. Go one level deeper. Within paid conversion, ask: what do users who convert have in common that users who don't convert lack? Is there a specific feature they've used? A usage threshold they've crossed? A company size or ICP profile that predicts conversion?
For engagement, look at your retention curve. Plot retention by cohort over 8 weeks. A curve that flattens into a baseline indicates you have a retained core. A curve that trends toward zero means the product isn't building habit, which is a much harder problem.
The most expensive bottleneck analysis mistake is optimizing the wrong stage.
Teams do this constantly. A marketing leader who understands demand generation improves acquisition. A product manager who owns onboarding improves activation. Nobody owns the engagement-to-paid transition, so it goes unfixed while everyone else's work flows through a broken gate.
If your engagement-to-paid rate is 14% and your activation rate is already at 55%, the next dollar of investment goes to paid conversion, not acquisition. More users at the top do not fix leakage at the bottom. They just make it more expensive.
The second version of this mistake: fixing the symptom rather than the cause. Engagement looks low, so you add push notifications. But engagement is low because the product never delivered value at activation. Users aren't coming back because they never found a reason to come back in the first place. Fix activation first.
Pull your four conversion rates for the last 30-day cohort. Write them down. Find the lowest one. That stage gets your next sprint.
If you don't have the instrumentation to pull these numbers yet, start there. Define your activation event, make sure it's tracked in Amplitude or Mixpanel, and build a cohort funnel. You cannot run a PLG funnel bottleneck analysis without this baseline, and running the analysis is the fastest way to stop guessing about where your growth is stuck.