processApril 11, 20242 min read

The useful few

Most of a product's use tends to gather in a small part of it. Finding that part is a measuring job, not a slogan.

Most products have a few features people use every day and a long tail they rarely touch. The "80/20 rule" is the familiar name for that shape. I find it useful as a pattern to go looking for, and misleading when it's treated as a law.

Feature usageWrite: 1000Search: 330Open recent: 172Share link: 109Tags: 76Pin: 57Export PDF: 44Templates: 36Shortcuts: 30History: 25Comments: 22Import: 19Backlinks: 17Calendar: 15Word count: 13Focus mode: 12Print: 11Themes: 10Attachments: 9API: 8

Top 4 of 20 features (20%) · 80% of use

Where the use actually isTwenty features of an imagined notes app with invented usage. The darker bars are the top fifth.bars travel 500ms, 12ms apart · heights tween 500ms · line fades in 300ms after · ease (0.22, 1, 0.36, 1)

Measure the curve you have

The study ranks twenty invented features by use and darkens the top fifth. With the focused dataset, those four cover 80% of the use. Switch to the spread-out one and the same four cover about 40%. Neither number is the right one. They describe two different products.

That's why I'd rather measure than assume: feature analytics, support tickets, crash reports, whatever shows where people actually spend their time.

Microsoft did this with crash reporting. In an October 2002 email, Steve Ballmer wrote that about 20 percent of bugs caused 80 percent of all errors, and that one percent of bugs caused half of them (quoted on Microsoft's archived engineering blog). That was their data, and it pointed them at the fixes that mattered most. It says nothing about anyone else's bugs.

A misnamed principle

The name is a small cautionary tale of its own. The quality engineer Joseph Juran called it the principle of the "vital few and trivial many" and attached Vilfredo Pareto's name to it. Pareto had studied the unequal distribution of wealth. In a 1974 essay, Juran admitted he had "mistakenly applied the wrong name to the principle." The label stuck anyway.

I like the story because it shows a useful observation turning into a slogan. The observation is that effects are often concentrated. The slogan says it's always 80/20.

What I do with it

When I'm deciding where to spend design time, I look for the small set of screens and controls that carry most of the use, and I polish those first: their states, their speed, their words. The long tail still has to work, because someone's important task lives there. It just waits its turn for polish.

Then I check the curve again a few months later. Products change, and the useful few move with them.