Most UX advice is taste with a vocabulary. This is not that. I am an Industrial and Organizational Psychologist who also writes the code, which means every recommendation comes with a measurement, a reason from the research, and a number you can check yourself.
A hundred-item checklist is a way of avoiding the decision about what matters. Each of these fourteen can be measured rather than argued about, and each has a cost to the reader that fits in a sentence.
Three kinds of load, only one worth removing. Some simplification makes a product measurably worse, and knowing which is the job.
Every choice you put in front of somebody has a price. The page that converts is usually the one that asks for fewer things, not the one that argues better.
Placement matters more than the signal. A guarantee that appears after the payment field has already been skipped.
Most of the traffic and most of the failures. Five things get worse on a phone and only one of them is screen size.
A page can be indexed and still never be quoted. Increasingly the first answer somebody sees is not a link at all.
Where the method comes from, the fourteen checks, and what the work looks like.
There are two kinds of people who give advice about websites. Designers, who know what looks right and are often correct without being able to say why. And marketers, who know what converted last time on a different site with a different audience. Both are useful. Neither can tell you what a page is doing to the person reading it.
That question belongs to psychology, and it has answers. How much a person can hold in working memory. What happens to a decision when you add a seventh option. Why a form with eleven fields collects fewer completed forms than one with four, even when nobody complains about the eleven. These are measured findings, not opinions, and they have been measured for decades.
UX psychology is not a separate discipline with its own literature. It is Industrial Psychology pointed at an interface instead of at an organisation. The field studies how people behave inside systems designed by other people: what they notice, what they misread, how much they will hold in mind before giving up, and what a badly designed process costs the person going through it. A checkout is that, and so is a job application form.
My degree is in Industrial and Organizational Psychology, which is the full name of the same field. The Industrial half is the part that transfers directly to a web page, and it is why the recommendations here come with measurements rather than with taste.
These are the checks behind the free tool on this site. You can run it on any address in about thirty seconds and see your own results before you ever speak to me.
Two of them will usually not apply to your page, and the tool says so rather than quietly scoring them as passes. An earlier version scored a blank page at 75 out of 100 because it had no images and no form, so “0 of 0 images have no alt text” collected full marks twice for content that was not there.
You know what it says, so you can read it at any contrast. Your visitor is reading it for the first time, often outdoors, often on a phone at 40% brightness. This is the single commonest failure and the cheapest to fix.
Someone scanning a page reads the links and almost nothing else. Eleven identical links tell them nothing, and a screen reader announces the same meaningless phrase eleven times.
A second request the visitor feels as a pause between the tap and anything happening. Nothing is broken, which is exactly why it survives for years.
It starts with measurement, not opinion. I run the fourteen checks, then I read the page the way somebody who has never seen it reads it: five seconds, then the squint test, then the links on their own with nothing else visible. Those three cost nothing and they repeatedly find that the most prominent thing on a page is not the thing the page is for.
Then the findings get ordered by what each one costs the reader, not by how easy it is to change. That ordering is the part clients tell me they had never seen before, and it is the difference between a list of twenty problems and a decision about what to do on Monday.
You get a written report, and it is written to be handed to somebody else: your developer, your agency, your co-founder. A finding that is correct and delivered badly changes nothing, so there is a section on how to say each one in a meeting.
No dark patterns. Not because it is a nice position to hold, but because regulators now enforce against several of them and the fines are real. Confirmshaming, fake scarcity counters, pre-ticked consent, subscriptions that take four screens to leave: these convert in the quarter and cost more than they earn over any longer period.
And no promise about your conversion rate. These checks measure a page's mechanics. A perfectly audited page selling something nobody wants will still not convert, and any consultant who promises you a percentage before seeing your offer is selling you something other than analysis.
Most people writing about UX psychology have one half or the other. They can quote Kahneman and cannot open a stylesheet, or they can ship a site and have read a blog post about Cialdini. Holding both is why the examples on this site are measurements from real work rather than principles quoted from a book: a contrast fix that measured as not having happened because a stylesheet was re-declared later in the page, a layout shift confidently predicted and measured at zero, and one page that turned out to be 89% advertising script by weight.
Do not take my word for any of this. Run the fourteen checks on your own page and read the results yourself.
Fixed scope and a fixed price. No hourly billing, no retainer you have to escape from.
One page, measured and written up
Up to 8 pages, plus the journey between them
I write the code as well
What this is, what you get, and what it will not do.
No. A UX designer decides what the interface should look like and how it should behave. I measure what it does to the person using it, and say what that costs. The two work well together and they are not the same job — I am often brought in alongside a designer rather than instead of one.
No, and be careful with anybody who does before seeing your site. These checks measure the mechanics of a page. If your offer is wrong, or you are buying the wrong traffic, every check here can pass and the page will still not convert. The audit will tell you that too, which is worth knowing.
A written report: the score, what to fix first, every check in full with the number it found and what that number costs the reader, and how to say each finding to whoever has to act on it. It is written to be forwarded.
No. The audit tool on this site runs all fourteen checks on any public page for free, as many times as you like, with no email required. Forty-seven full audits of well-known brands are published here in full as well. Buy the service when you want the parts a tool cannot do.
Yes. The work is remote and the psychology does not change with the country. Payment by Stripe or PayPal, in USD.
Only the kind that survive scrutiny. Removing friction, making the next step obvious, and putting reassurance where doubt actually occurs. Not fake counters, not confirmshaming, not consent that is hard to withdraw — those are now regulated in several markets and they cost more than they earn.
Run the free tool first. If the results raise questions, that is when to talk.
Your cart is empty
Browse Shop