Innovation by Subtraction

What it is

Adams and her colleagues set people a range of problems, some physical, some abstract, some structural, in which a solution could be reached either by adding an element or by removing one. Across eight experiments, people reached for addition by default. Subtraction was considered mainly when it was explicitly prompted, or when participants had enough time and enough cues to reconsider their first approach.

The finding is about search rather than preference. It is not that people evaluate removal and reject it. It is that removal does not reliably enter the set of options being evaluated at all, so the additive solution wins an election it ran in unopposed.

A replication and extension by Fillon and colleagues in the Journal of Creative Behavior in 2025 confirmed the core effect. Later work finds it varies with age, culture and task, which locates the boundaries rather than undermining the result.

The lineage is worth stating because the label obscures it. Klotz popularised the research in his trade book Subtract, also in 2021. McNeill's phrase is a label attached to an existing body of work, and it should be credited as a label only.

In effect

The ordinary case is a checkout page with a falling conversion rate. The team proposes a trust badge, a live chat widget and a progress bar. The fix that works is deleting four optional form fields. Nobody suggested it first, because nobody thinks of less as a feature.

The organisational version is worse than the individual one, because institutions reward visible addition. A new process is a deliverable. A deleted process is an absence, and absences do not appear in a performance review.

What follows practically is procedural rather than attitudinal. Since the bias is a failure to generate the option, exhortation does little and a prompt does a lot. Requiring a subtractive candidate alongside every additive one, or asking explicitly what could be removed before asking what could be added, puts removal into the comparison where it can win on merit.

What it does not say

It does not say that less is always better. The finding is that subtraction is under-considered, not that it is usually right. A prompted comparison can perfectly well conclude that the addition wins.

It does not say people dislike subtraction. They overlook it, which is a different failure and has a different remedy.

It does not quantify a business cost. The experiments measure how often people find the subtractive solution, and the translation into money saved in any real organisation is not something the research supports.

And it does not belong to whoever named it most memorably. The evidence is Adams, Converse, Hales and Klotz's.

---