Adding a feature can make your product look worth less

Suppose you’re selling a dinnerware set. Option A: 24 pieces, all intact. Option B: those same 24 pieces, plus 16 more — except a few of the extras are chipped or broken. Strictly, B is more dinnerware; it contains everything in A and then some. Yet when people are shown only one of the two — not both side by side, just one — they’ll often pay more for the smaller, all-intact set A than for the larger set B with the broken pieces in it. Adding cups and saucers to the offer made buyers value it less.

This is Christopher Hsee’s work, and the counterintuitive result has an equally counterintuitive name: the less-is-better effect, which falls out of a deeper idea he called the evaluability hypothesis. The key move is noticing how the buyer is judging. Shown a single set with no comparison in front of them, they have no idea whether 24 pieces or 40 is “a lot” — number-of-pieces is a hard attribute to evaluate in isolation. But “are the pieces broken?” is dead easy to evaluate with no reference at all. So the easy-to-judge cue takes over the whole valuation, and a set that reads as “has broken things in it” loses to one that reads as “everything’s intact,” even though the first is objectively more stuff. When you can’t judge the important number on its own, you fall back on whatever cue you can judge — and let it stand in for the whole.

The engine underneath is the difference between two modes of judging, and it’s the most useful thing here to hold onto. In joint evaluation — two options side by side — the comparison itself makes the hard attribute legible: 40 next to 24, now you can see which is bigger, and the broken pieces matter less because the sheer quantity is finally visible. In separate evaluation — one option, alone — the hard-to-assess attributes go quiet and the easy-to-assess ones run the show. The same offer can win or lose depending only on which mode the person is in. And here is the fact that should make every builder sit up: your user is almost always in separate evaluation. They are looking at your product, on your page, in their life — not holding it up against a competitor in a lab. They are judging you alone.

⁂

Once you see the product through that lens, a pile of scattered decisions turns into one. When you add a feature, a plan tier, a settings toggle, a headline claim — you are adding another attribute to a thing the user is judging alone. If the addition is easy to evaluate and reads well, great. But if it introduces something the user can easily judge and judges poorly — a half-baked feature, a confusing extra plan, a “pro” tier whose value is illegible — it can drag the whole perception down, exactly like the broken cups. More surface area is more chances for an easy-to-read flaw to become the cue that stands in for everything. This is the mechanism under “we added it and conversion dropped.” The feature wasn’t neutral. In separate evaluation, its most-evaluable quality got promoted to the summary of the entire product.

It also reframes a choice you make constantly without noticing: do you show options together or one at a time? A comparison table puts the user in joint evaluation, where your genuinely-bigger, genuinely-better spec becomes legible — good, if the truth is on your side. A single landing page with one plan keeps them in separate evaluation, where a clean, complete-feeling offer beats a bigger-but-messier one, and where hard-to-judge magnitudes (“2,000 credits”) do little unless you translate them into something evaluable (“about a month of daily use”). Neither mode is honest or dishonest by itself; the point is that you’re choosing the mode, and the mode decides which of your attributes the user’s judgment actually runs on.

⁂

The caveat is the same one that guards every perception lever: this is about perceived value, and the honest use is to make real value legible — not to hide real value behind a flattering frame, and certainly not to strip a product down and call the loss elegance. “Less is better” is a fact about evaluation, not a mandate to ship less. Sometimes the right response is to cut the illegible extra. Sometimes it’s to keep the extra and do the harder work of making its value easy to judge. What you shouldn’t do is keep adding evaluable flaws and wonder why more product keeps reading as less.