UX Tools UX Teardown: A Table That Shows Its Sample Size and Still Misleads

Most comparison tables on the internet hide their arithmetic. UX Tools doesn't. Its Design Tools Database prints the response count in parentheses after every figure: Figma at 82.3% usage, and then, quietly, (1620). Four rows down, that parenthetical is also what makes the table's biggest problem visible to anyone reading carefully and invisible to everyone else.

This teardown walks the three surfaces UX Tools builds from one survey and names where the layout stops keeping the promise the data makes.

Three views, one dataset, one rail

The browse experience is a single structural idea repeated three times: a filter rail on the left, results on the right, no other chrome. Categories on the database, topics on the articles, a numbered contents list on the survey. Step through it:

Consistency is the fourth of the ten heuristics (NN/g), and this is what it buys when you take it seriously. Once you've used the tools database, the articles index costs nothing to learn. The rail is where the rail was, the active item is the same pink block, the results reflow in the same grid. You pay for the pattern once and collect on it across the whole site.

The parenthetical is the best thing on the page

Look at what each row actually carries.

Two metrics, each with its n: usage 82.3% (1620), rating 4.57 (1603). Almost every tool roundup you've read publishes the first number and buries the second, because "4.6 stars" reads better than "4.6 stars from eleven people." Printing n is what separates a ranking from a measurement. NN/g's work on data tables is blunt about the underlying task: users come to a table to compare, and comparison needs the units to mean the same thing in every row (NN/g — Data Tables).

Which is exactly where this table breaks its own rule.

Sorted by usage, the list runs Figma (82.3%, n=1620), Sketch (1.8%, n=35), Adobe XD (1.4%, n=28), and keeps descending until it reaches Balsamiq: 0.04% usage, one response, rating 5.00. That 5.00 sits in the same column, in the same type size, in the same shade of green as Figma's 4.57. One row averages 1,603 opinions. The other is a single person's click. The data is honest; the layout asserts an equivalence the data never claimed.

The disclosure is already there, so more of it won't help. What's missing is confidence rendered as weight: dim or suppress the rating below a threshold, or let n govern how loud the row is allowed to be. Right now the parenthetical is a footnote asked to do a layout's job, and footnotes lose that argument every time (NN/g — Comparison Tables).

The awards live at /survey/design-tools-awards

The award pages are where the dataset becomes an argument, and they're written to be quoted.

Design Standard Award, winner Figma, three stats laid out like a verdict: 82.3% market share, a 46:1 ratio against the nearest competitor, 93.1% adoption among corporate ICs. The 46:1 is the sharpest thing on the whole site. It takes two numbers already sitting in the table and turns them into a sentence you'd repeat in a meeting. A numbered [001]–[010] rail on the left and a PREV / NEXT pair up top thread nine awards into something you read start to finish.

The problem is finding it. These pages live at /survey/design-tools-awards/<award>, inside a report, behind a dropdown. There is no /awards. The nav's top-level items are Design Tools, Resources, Articles, Episodes; the awards are a chapter of a document reached through the third one. That's an information-scent failure (NN/g). The label a visitor hunts for doesn't appear at the level they're hunting at, so the most citable artifact here is the hardest one to arrive at cold.

The freshness gap nobody labels

The award page and the database both draw on the same report, and the report is the 2024 survey. It says so in the sidebar mark and again in the footer nav. The articles beside them are dated May 2026. The banner across the top promotes a newer State of Prototyping survey.

None of that is hidden. None of it is stated where it matters either. The database's own subhead reads "See what's rising, falling, and what designers actually rate highest" — present tense, live-feeling language, over a two-year-old snapshot. A reader arriving from search sees current-year articles in the same visual system as the numbers and has no reason to date the numbers differently. This is visibility of system status applied to the data rather than the interface (NN/g): the system knows when it was measured and doesn't say so at the point of comparison. One line above the table would close it. Data: 2024 Design Tools Survey.

A grid pretending to be a list

Every article card on that index wears the same generic UX Tools logo art. The image differentiates nothing, so the tag and the date carry all the scanning work while the thumbnail eats the vertical space. The same content set as a dense list would let you scan twice as far in the same screen.

The homepage sells the newsletter

Scroll the full page and the priority is unambiguous.

"The Strongest Signal in Design Tools," a subscribe field above the fold, JOIN 90K+ READERS FROM with six company marks, then a long stack of evidence for that one claim: new releases, podcast episodes, articles, a testimonial wall that names each reader's role and employer, an archives list at the bottom. Every module argues the same proposition, and the dark, high-contrast canvas keeps a very long page feeling like one continuous case.

What it never does is show a number. The database is the one asset here nobody else has. The awards are built on it, half the articles cite it, and on the homepage it appears as a nav item. A page this committed to proving signal could prove it with one row of the table it already owns.

What this means for your product

Publish your n. If you rank, rate, benchmark, or score anything, the sample size belongs next to the number, and the fact that it makes some rows look weak is precisely the point.

Take it further

Whether a number tells you what it means, what it rests on, and whether it's comparable to the one beside it is a clarity question. It's the same lens as the UX Clarity framework, which is what we run in a Full UX Audit. For how that scoring turns into prioritized fixes, read what a real UX audit looks like.

Sources: NN/g — 10 Usability Heuristics · NN/g — Data Tables: Four Major User Tasks · NN/g — Comparison Tables · NN/g — Information Scent · NN/g — Visibility of System Status.

Ready to find out whether your own numbers survive the layout you put them in? Apply for a Full UX Audit →