Score Your Signup Flow Out of 18 Before You Redesign Anything

Most teams respond to a soft signup number the same way: they redesign. New hero, new illustration, new button color. Three weeks later the number hasn't moved, and nobody can say which of the twenty changes was supposed to do it.

A teardown is cheaper and tells you more. You score the flow you already have against a fixed set of heuristics, find the two or three moments that actually cost you signups, and fix those. Redesign is what you do after a teardown comes back clean.

Here's the version we run on signup flows, scored out of 18.

Why a signup teardown beats a redesign

A redesign is unfalsifiable. When the number moves afterward, you can't attribute it. A scored teardown gives you a before/after on one specific moment.
Redesigns fix the wrong layer. Founders look at a "boring" page and assume design. Most signup leaks are field count, unclear value, and doubt at the moment of the ask — none of which a new font solves.
A score creates a queue. "Signup feels off" isn't a task. "Trust scores 0 because the reassurance line is in the footer" is.
It's fast. Thirty minutes, one incognito window, one phone.

The goal isn't a prettier page. It's finding the exact moment a stranger decides not to continue.

The 9 heuristics: score your signup flow out of 18

Score each one 0, 1, or 2. Zero means missing or actively working against you. One means it's technically there but a stranger wouldn't notice it. Two means it's clearly handled without effort.

Walk the flow as someone who has never heard of you. If you can't unsee your own product, hand your phone to a colleague and watch their face instead.

Value clarity. In five seconds, can a stranger say what they get and who it's for? A headline that says "Get started" is a 0. One that names the outcome and the audience is a 2.
Field justification. Every field has to earn its place before the first session. Company size, phone number, "how did you hear about us" — cut it or defer it until after they're in.
Effort preview. Do they know how much work this is before they commit? "Step 2 of 3" or "takes about 90 seconds" turns an unknown into a known.
Trust at the point of risk. Doubt spikes exactly where you ask for an email, a password, or a card. Reassurance belongs next to that field, not in the footer under a copyright line.
Action hierarchy. One primary button. If Google, email, and "Book a demo" all look equally loud, you've handed the decision to the visitor.
Validation feedback. Errors appear inline, on blur, in plain language, and never wipe what they typed. Catching email domain typos as they type is a free 2.
Recovery paths. "Already have an account?", a magic-link fallback, a resend button that actually resends. Signup shouldn't be a one-lane road with no exits.
Post-submit expectation. After they hit the button: what happens, where do they look, how long does it take. Silence here is where a lot of "drop-off" really lives.
Evidence near the ask. One line of proof — a customer name, a specific claim, a recognizable logo — inside the same visual block as the button. Proof three scrolls up doesn't count.

Add it up:

16–18: the flow isn't your problem. Look at traffic quality or first-session activation.
11–15: fix the 1s. You're leaving real points on the table, but nothing is broken.
6–10: there's at least one P0 in here.
0–5: stop touching your landing page.

If you'd rather not score it by hand, run a free audit on your signup flow and compare the output against your own scoring. The disagreements are the interesting part — they usually mark the heuristics you're too close to see.

Before and after: the same form, 7/18 to 15/18

This is a compressed version of a pattern that shows up constantly — a B2B tool with a five-field signup and zero trust copy anywhere near the button.

Before (7/18):

Create your account
[Work email]
[Password]
[Company name]
[Phone]
[ ] I agree to the Terms and Privacy Policy
[Get Started]

Where it lost points: value clarity 1, field justification 0, effort preview 0, trust 1, hierarchy 1, validation 2, recovery 1, post-submit 1, evidence 0. Two of those fields were never used before week two, and the only trust signal was a link buried inside a checkbox label.

After (15/18):

Start your 14-day trial
No card. Cancel from settings in two clicks.

[Work email]            We'll send your login link here.
[Continue with Google]  [Continue with Microsoft]

[Start free trial]      Next: name your first project (about 90 seconds)

Already using Acme? Sign in

What changed, and why it scored higher:

The headline names the deal. "Start your 14-day trial" beats "Create your account" because it says what they get, not what they're doing for you.
Two fields instead of five. Company and phone moved into onboarding, after the first value moment. Field justification goes 0 to 2.
Reassurance sits under the button, not in a footer. "No card" and "cancel in two clicks" are the two objections that actually stop people.
One primary action. OAuth becomes a real alternative instead of a competing button.
The next step is previewed. "About 90 seconds" plus a named next screen kills the blank-page anxiety that follows submit.

Still 15 and not 18: no magic-link fallback