Why this matters
Conversion rate is where the smoke test stops being a feeling and becomes a number. But the number only means something if the threshold was set BEFORE looking, and if traffic is segmented by source, otherwise founders explain away weak results and call strong noise a trend.
What "done" looks like
- A pre-registered threshold written down before traffic ran (e.g. strong ≥10%, weak <3% on targeted traffic)
- Conversion computed per source, not blended
- Enough visitors per source (~100) that the rate is not pure noise
- A written verdict: strong / mixed / weak, and what you will do next
How to do it
- Write the thresholds first. Targeted warm traffic to a waitlist typically converts 10 to 25% when the pain is real; cold paid traffic 2 to 5%. Set your lines, then look.
- Segment ruthlessly: community traffic, outreach traffic, and paid traffic are different tests. A blend hides the answer.
- Exclude the sympathy vote, friends, family, fellow founders. They convert out of kindness.
- Read the thank-you answers ("what made you sign up?") alongside the rate; five consistent answers explain the number.
- Decide per the pre-set rule, and record the decision, this is evidence for the go/no-go call ahead.
Common mistakes
- Moving the goalposts after seeing the data ("well, 2% is fine actually")
- Judging conversion on 20 visitors, that is a coin flip, not a rate
- Optimizing button colors when the problem is the pitch or the audience
Real-world examples
- The discipline that makes a signup test honest is setting the pass/fail threshold before the numbers arrive. Decide in advance what rate would count as a green light, because after the fact almost any result can be rationalized as encouraging.
- Track the whole funnel, not just the headline rate: visitors → signups → activation. A strong signup percentage on the wrong traffic tells you little, which is why it pairs with driving real visitors to the page and proper analytics.
- Treat the number as an input to the go/no-go decision, not the decision itself, conversion tells you interest, but you still confirm intent with conversations or pre-orders.
From a founder's point of view
The trap here is measuring after you've already fallen in love with the idea, at which point the number is just something you interpret to feel good. Writing the threshold down first is a small act of intellectual honesty that saves months. A clear, pre-committed line turns a fuzzy "seems okay?" into an actual signal you can act on.
Rule of thumb
Decide what the number must be before you see it. If you find yourself explaining why a low number is "actually fine," the test already gave its answer.