Picture a sustainability analyst who opens a big spreadsheet every morning to hunt for something specific, like a number that doesn’t add up or a supplier who’s gone quiet.
Now someone "modernises" it into a dashboard, with summary tiles on top and cards instead of rows. It looks cleaner and friendlier, and it’s slower to use.
UX has a healthy obsession with reducing cognitive load, and the research behind it is real. Miller’s famous 1956 paper on the limits of working memory has shaped how designers think about density for decades. The catch is the word unnecessary. In a consumer app, most complexity is unnecessary, and simplifying makes things better. In a professional analytics tool, a lot of the complexity is there because the work is complex and the people doing it are experts.
Remove it and the load just moves somewhere else, usually into three extra clicks to rebuild the picture one good table would have shown them.
There’s a name for part of this. Pirolli and Card called it information scent: the cues that tell people they’re getting closer to what they need. Summary cards and collapsed sections hide those cues, so experts spend their time hunting instead of working. The screen looks simpler while the person using it does more of the thinking.
At Scaler I ran evaluation sessions comparing table densities, using realistic datasets instead of a tidy demo version. The sparsest layout, with the most whitespace and the most "designed" look, had the lowest accuracy on column-scanning tasks and needed the most scrolling. The densest one was bad for scan speed. A medium-high density with deliberate spacing rules won on everything we measured.
Medium-high density was the fastest to scan
The same emissions data in the same space
Scope 1 and 2 emissions, tCO₂e.
Spacious, 40px rows, ruled out
Looked the nicest, but people scrolled forever.
| Facility | Scope 1 | Scope 2 |
|---|---|---|
| Rotterdam warehouse | 1,284.6 | 2,310.0 |
| Lyon office | 42.1 | 188.4 |
| Gdańsk plant | 3,907.3 | 5,122.8 |
| Leeds distribution centre | 611.0 | Missing |
Showing 4 of 847 facilities
Medium-high, 32px rows, chosen
Fastest to find a row, and the most accurate when scanning across columns.
| Facility | Scope 1 | Scope 2 |
|---|---|---|
| Rotterdam warehouse | 1,284.6 | 2,310.0 |
| Lyon office | 42.1 | 188.4 |
| Gdańsk plant | 3,907.3 | 5,122.8 |
| Leeds distribution centre | 611.0 | Missing |
| Porto office | 18.9 | 96.2 |
Showing 5 of 847 facilities
Maximum, 24px rows, ruled out
Most data on screen, but scan speed fell apart.
| Facility | Scope 1 | Scope 2 |
|---|---|---|
| Rotterdam warehouse | 1,284.6 | 2,310.0 |
| Lyon office | 42.1 | 188.4 |
| Gdańsk plant | 3,907.3 | 5,122.8 |
| Leeds distribution centre | 611.0 | Missing |
| Porto office | 18.9 | 96.2 |
| Brno data centre | 74.5 | 4,480.1 |
Showing 6 of 847 facilities
Experts and beginners aren’t the same user, and research on expert analysts points the same way: hiding data in the name of simplicity makes experts slower and less accurate, even when it helps beginners.
In sustainability work, the details that get hidden first are often the ones that matter most. Whether a figure is Scope 2 or Scope 3, or whether a supplier number was verified or self-reported, can decide whether a compliance filing is right.
None of this means simplifying is wrong. Decoration and pointless repetition are noise, and they should go. Alberto Cairo argues in The Truthful Art that the job is to make complexity navigable, not to pretend it isn’t there, and I think that’s the right test for any expert tool.
The tile says what the number is
Supplier emissions, Q4 2023
42.3 tCO₂e
▼ −14%, lower, which is good vs 49.2 in Q4 2022, recalculated
Restated: the method changed in Q3, so earlier quarters were recalculated.
Illustrative data
The context is behind See details
Supplier emissions, Q4 2023: details
Supplier emissions, Q4 2023
42.3 tCO₂e
▼ −14%, lower, which is good vs last year
Illustrative data
So before cutting anything, I ask whether it’s noise or signal. Simple means fewer things on screen. Legible means the right things, arranged so people can find them and act on them. For expert work, those two often pull in opposite directions.
References