The method · Free · 10 a week
The Uxerfy Score, and how I arrive at it.
Every audit ends in one number between 1 and 10. This page is what is behind it: the four lenses each page goes through, what the bands mean, and how a finding ends up as a critical blocker rather than a quick win.
The score is a judgement, not a calculation. The four lenses decide what gets examined and in what order; the number is my read of the page as a whole, weighted by what a visitor actually loses.
I could publish a weighting and multiply four numbers together, and it would look more rigorous than it is. What makes this checkable is not arithmetic — it is that the same lenses run on every page, that every finding says what I saw and why it costs you, and that the teardowns put the whole reasoning in public where you can disagree with it before you send me anything.
The four lenses
Every page goes through these, in this order. The order matters: a page that fails the first lens rarely gets a fair reading of the fourth, because nobody stays long enough to reach it.
-
First impression
What a first-time visitor understands in the first five seconds, and whether the page offers one obvious next step or several competing ones. Most drop-off happens here — before anyone reads a word of your feature list.
- Can someone say what this is, in their own words, after five seconds?
- Is there one obvious next action, or several competing for the same click?
- Does the first screen show the thing the visitor came for?
-
The path to value
Every step between arriving and the core action: how many there are, which could be deferred, and where a form asks for more than the stage warrants. This is where signups leak quietly.
- How many steps stand between arriving and the core action?
- Which of them could be deferred until after the visitor is committed?
- Does any step ask for more than it has earned at that point?
-
Clarity of copy
Whether the page is written for the reader or for the system behind it. Buttons that do not say what they do, headlines that name a category instead of a benefit, errors that do not say how to recover.
- Does every button say what happens when you press it?
- Does the headline name a benefit, or a category?
- When something goes wrong, does the page say how to recover?
-
Trust and friction
Whether the price is legible, whether reassurances appear at the moment of hesitation, and whether anything asks for more commitment than the visitor is ready to give.
- Can a visitor work out what this costs without leaving the page?
- Do the reassurances appear where the doubt happens, or in a footer?
- Is anything asked for before the visitor has a reason to give it?
The full question list behind the four is public — it is the UX audit checklist, the same one I work through on every page. For the ground beneath all of it, see what a UX audit is and what it cannot tell you.
What the number means
Three bands, and the band matters more than the decimal. The difference between a 6 and a 7 is a judgement call; the difference between a 4 and a 7 is a different kind of problem entirely.
Every page is read through the same four lenses and scored on one scale, so the number means the same thing on your page as it does on anyone else's. Most real product pages land between 5 and 7.
- Losing people A first-time visitor cannot tell what this is, or the path to the core action is broken. Fixing one thing here moves more than a redesign anywhere else.
- Works, leaks Nothing is broken and motivated people get through. The losses are the ones you never see: hesitation, comparison, quiet exits.
- Deliberate Someone has thought about the sequence, the copy and the moment of doubt. What is left is refinement, not repair.
Blockers and quick wins
Every report contains two lists. What separates them is the kind of change required, not the size of the problem — a blocker can be one word, and a quick win can be a whole section.
3 critical blockers
The three findings costing you the most, in that order. A blocker usually needs a decision before it needs a designer — what the page is for, which plan you want people on, what you are willing to ask for up front.
5 quick wins
Changes you can ship this week without a redesign or a sprint of engineering time. They are not the biggest problems on the page. They are the ones where the distance between knowing and fixing is shortest.
How severity is decided
Severity answers what a finding costs, never how hard it is to fix. Those two questions have different answers often enough that mixing them is how cheap, harmless changes end up at the top of a report.
-
A visitor who hits this is likely to leave, or to finish the task with the wrong expectation. Costs you the conversion outright.
-
Slows the decision or adds doubt at a point where doubt is expensive. Most people get through; fewer than would have.
-
Noticeable to anyone reading closely, and worth fixing when the file is already open. Not what is costing you this quarter.
What the score does not measure
A method is only useful if it says where it stops. These are outside it:
- Whether real users succeed. That needs testing with real users, and this is an expert review.
- Visual taste. A page can be plain and score well, or beautiful and score badly.
- Technical performance, except where load or layout shift is what the visitor is reacting to.
- Accessibility in full. There is a separate audit for that, against WCAG.
- Anything behind a wall I was not given a way through.
The method, applied in public
Every teardown runs this method end to end on a page you already know, with the score, the blockers and the reasoning out in the open.
- A signup form is the one place a product has earned nothing yet. Every question that serves the business rather than the visitor is paid for in people who close the tab — and the question about money is the most expensive one to ask first.
- Four undifferentiated columns hand the sorting work to the visitor. One recommended plan and a stated saving do that work for them — and a decision you make easy is a decision that actually gets made.
- A headline price is the number the whole relationship is measured against. If the real bill is usage-based, the card has to say so on the page — otherwise the first invoice becomes the moment trust breaks.
Questions about the method
Is the score calculated from the four lens scores?
No. The lenses decide what gets examined and in what order; the number is my read of the page as a whole, weighted by what a visitor actually loses. A page can score badly on one lens and still land at 8 if that lens is not what the visitor is deciding on. Publishing a formula would claim an arithmetic that does not happen.
Then how is it not just an opinion?
It is a judgement — I would not pretend otherwise. What makes it checkable is that every page goes through the same four lenses and the same published checklist, and every finding states what I saw and why it costs you. You can disagree with the reasoning rather than the number. The public teardowns let you check that reasoning before you send anything.
Would the same page get the same score twice?
Within about half a point, yes — the lenses and the band definitions are what keep it stable. Across a whole point, probably not, and anyone claiming otherwise for an expert review is overselling. The score is there to rank pages against each other and to make the direction of travel visible, not to be a measurement instrument.
Why three blockers and five quick wins, and not more?
Because a list of thirty findings gets read once and actioned never. Three is what a team can hold in its head and argue about; five is what fits in a week. Everything I find goes into the report, but the two lists at the front are the ones I would do first, in that order.
Does a higher score mean I do not need the audit?
Often the opposite. An 8 means the obvious repairs are done and what is left is the expensive kind — the hesitation you cannot see in analytics. A 4 usually has one broken thing that costs more than everything else combined, and finding it is quick.
Why trust this audit
Product designer · Barcelona
Uxerfy is a UX audit service run by one person: send a URL, get back a PDF within 24–48 hours with a score out of 10, three critical blockers and five quick wins. I do every audit myself and put my name on it — here's who I am and how I work.
Verify me on LinkedIn →Here's a complete sample report. No email required — download it and see exactly what lands in your inbox.
Download a sample report (PDF) ↓See what this method finds on your page.
Send a URL and you'll have the report within 24–48 hours.
Get a free audit →