Choice architecture
Every way of presenting options shapes the choice, and there is no neutral way.
Choice architecture is the set of design decisions that surround a choice: what the options are, what order they appear in, what is preselected, what is described first, and what happens if the person walks away. None of these can be left blank. Once you accept that, the question stops being whether to influence and becomes which influence you are willing to defend.
How it shows up in software
A pricing page is choice architecture. So is a settings screen, a plan picker, a sort dropdown and the order of payment methods at checkout. Most of these get designed once, by whoever built the screen, and then run for years carrying assumptions nobody wrote down.
Using it well
- List the choices your product makes people make, then audit each one for option set, order, preselection and exit path.
- Show the outcome, not the attribute. Translate 8GB into how many photos it holds, and a rate into what the person pays per month.
- Structure large option sets before trimming them. Filters, categories and a sensible sort usually beat cutting real choices.
- Name the architecture in the design review so the trade-off is discussed by the team instead of set by whoever wrote the component.
Where it turns manipulative
- Decoy options exist to make a target look reasonable rather than to be chosen. If nobody is meant to buy the middle tier, you are steering, not offering.
- Ordering options by margin while implying the order reflects fit is a lie told with layout.
- Claiming neutrality is its own misuse. A team that says the design is just how it came out has stopped taking responsibility for an influence it is still exerting.
Where you have seen it
Stripe pricing page
Two clear paths, pay as you go and custom, with the fee arithmetic shown per transaction instead of as an abstract rate.
Apple iCloud storage picker
A small tier ladder with each tier described in what it holds, and the current usage shown next to it so the mapping from option to outcome is visible.
Airbnb search filters
Thousands of listings stay available, and the architecture is filtering and sorting rather than a shortened list.
What the research says
- Thaler, Sunstein and Balz, 2013 (Behavioral Foundations of Public Policy)Well evidenced
Sets out the working tools of choice architecture: defaults, expecting error, feedback, mappings from options to outcomes, structuring complex choices, and incentives.
This is a framework, not an experiment. Its value is as a checklist, and each tool has its own and much more variable evidence base.
- Johnson et al., 2012 (Marketing Letters)Well evidenced
A review of choice architecture tools separates those that structure the option set from those that describe the options, and catalogues how each changes what people pick.
- Iyengar and Lepper, 2000 (Journal of Personality and Social Psychology)Contested
The jam study: a display of 24 jams drew more browsers but fewer purchases than a display of 6, which became the standard citation for choice overload.
Scheibehenne, Greifeneder and Todd (2010) meta-analysed 50 experiments and found a mean effect near zero, with the outcome depending on prior preferences, expertise and how the options are organised. Treat too many options as a hypothesis to test, not a law.
Grades are a judgement about the evidence, not about how useful the idea is. Plenty of contested effects are still worth knowing, as long as you do not cite them as settled.