Ask a builder how their launched product is doing and you'll get a traffic number. Ask about their funnel performance — what fraction of visitors sign up, what fraction of signups reach the product's value, what fraction of those pay — and you'll usually get silence. That silence is expensive. "I get traffic but no sales" is the most common complaint in indie hacking, and it's never actually 1 problem: it's 4 possible problems wearing the same trench coat, and without funnel numbers you can't tell which one you have. Which means you can't fix it, so you do what builders do — ship another feature — and nothing changes. This post covers the 4-stage funnel every product has, what to instrument, what rough good and bad look like, and how to find your leak this week.
The only 4 numbers that matter early
Every product, from a CLI tool to a B2B SaaS, has the same skeleton:
Visit → signup → activate → pay.
Visit is someone landing on your page. Signup is an account. Activate is the first moment of real value — the file converted, the project created, the report generated. Pay is money. Between each pair of stages, people leak out, and the 3 leak rates between those 4 stages describe your entire business more honestly than any traffic graph.
Activation is the stage builders skip, and it's the one that matters most. A signup who never reaches value isn't a user — they're an email address with regrets. Define activation concretely for your product before you instrument anything: not "logged in twice" but "completed the core action once." If you can't name your product's activation moment, that's your first finding.
Instrumenting it: an afternoon, not a sprint
You need 4 events and a tool that draws funnels. PostHog is the default answer for indie products — its free tier comfortably covers a new launch, and the funnel view is built in. Plausible plus custom events works for the privacy-minded; Amplitude if you outgrow both. The tool genuinely matters less than the discipline.
Fire visit on landing (your analytics does this automatically), signup on account creation, activated on the first completion of your core action, and paid on the first successful charge — from your Stripe webhook handler, not the client, because client-side purchase events get eaten by ad blockers and you'll undercount the only number that matters. Add 1 property everywhere from day 1: the referrer or UTM source. Same 4 events, split by source, is how you'll later learn that Reddit traffic activates at 3x your X traffic — the single most actionable cut you can make.
Then build the funnel view and put it somewhere you'll see weekly. Numbers you don't look at are numbers you don't have. If you're pre-launch, do this before launch day — it's step 3 of how to launch your MVP for a reason. Launching without it means your launch teaches you almost nothing.
What good and bad look like — hedged, because context is everything
Benchmarks vary wildly by product, price, and traffic source, so treat every number here as a rough prior, not a target. With that hedge stapled on:
Visit → signup: for a free-signup product, something in the low single digits to ~10% is a common range; under 1% usually signals a landing page problem. For products asking for a card upfront, far lower is normal.
Signup → activate: commonly somewhere in the 20–50% band for products with decent onboarding; if fewer than 1 in 5 signups ever reach value, your onboarding is the fire to fight.
Activate → pay: the widest range of all — freemium products often convert low single digits of activated users; trials convert much higher. Anchor on your own baseline rather than anyone's blog post, this one included.
The honest takeaway isn't the specific numbers — it's the shape. A healthy funnel leaks gradually. A broken funnel has 1 stage that falls off a cliff, and that cliff is almost always visible within your first few hundred visitors. You don't need statistical significance to spot a stage converting at a tenth of the others.
Finding the leak: the diagnosis tree
Once instrumented, "traffic but no sales" resolves into 1 of 4 specific problems:
Leak at visit → signup: your page isn't converting the traffic you get. Either the traffic is wrong (check conversion by source — 1 bad channel can drown the average) or the page is unclear: vague headline, no price, weak CTA. Fix the message before the product.
Leak at signup → activate: people want it, then get lost. Watch 5 session recordings of fresh signups (PostHog includes this) — you'll usually spot the wall in the first 3. Common culprits: an empty state with no next step, required setup before any payoff, a core action that's hidden.
Leak at activate → pay: they get value and won't pay for it. This is a value-or-pricing problem — the free tier may give away the whole meal, the price may mismatch the wallet, or the upgrade moment may never surface. Tactics in pricing your first product.
No leak — just no volume: the funnel converts fine and 30 people saw it this month. That's not a funnel problem; that's distribution, which is a different post and a different fix.
One discipline turns this from a dashboard into a system: change 1 thing per leak, per week, and check the stage rate after. Shipping 5 fixes at once means learning nothing from whichever one worked.
The uncomfortable part: operating the funnel is a job
Here's what nobody tells builders: instrumenting the funnel is the easy 10%. The other 90% is operating it — checking the numbers weekly, forming a hypothesis about the worst stage, shipping the 1 change, reading the result, repeating. Forever. It's the least builder-shaped work that exists: no compiler, no green tests, just slow noisy feedback loops. It's also where products actually get good, which is exactly why so many technically excellent projects stay unlaunched or stall at a trickle — the building rewarded you, and the operating doesn't.
That operating loop is the core of what LaunchBuddy does. It's a launch studio: submit your unlaunched project, and if we pick it, we build it onto our harness — the 4-event funnel instrumented from day 1, on rails we've built once so they work every time — launch it live, and run that weekly diagnose-fix-measure loop as the job, not the chore. You keep ownership, pay a flat fee or share revenue, and can kill or port out anytime, taking the funnel data with you.
If you've got a project with traffic and no sales — or no traffic and no idea why — get the free honest assessment. If it's a no, you get the why, stage by stage. 60 seconds to submit at launchbuddy.app.