The paradox of choice

In short

If your shelf holds 24 flavours of jam, more people stop, but fewer may buy. The paradox of choice is the name for this contradiction. It is also called choice overload.

The idea: many options look attractive at first. Then the effort of comparing grows. The decision gets postponed or isn’t made at all. And even if it is, satisfaction with the choice can drop.

The effect is disputed. It doesn’t show up on every full shelf. Below I describe, with two meta-analyses, the conditions under which it occurs.

The jam experiment (2000)

The experiment was run by Sheena S. Iyengar (Columbia University) and Mark R. Lepper (Stanford University). The paper appeared in 2000 in the Journal of Personality and Social Psychology. Its title: “When Choice is Demotivating: Can One Desire Too Much of a Good Thing?”.

The experiment took place at Draeger’s grocery store in Menlo Park, California. According to the paper, the store sold over 300 varieties of jam. On two consecutive Saturdays, a tasting booth with Wilkin & Sons jams was set up there.

The booth showed either 6 or 24 flavours, and the display was rotated hourly. Flavours everyone knows, such as strawberry and raspberry, weren’t in the selection at all. Everyone who came to the booth got a $1 coupon; a code on the coupon showed which booth the person had seen.

Results of the 24-flavour and 6-flavour booths in the jam experiment

Measure	24 flavours	6 flavours
Passed the booth	242	260
Stopped	145 (60%)	104 (40%)
Bought, of those who stopped	4 (3%)	31 (nearly 30%)

Source: Iyengar and Lepper (2000), study 1. Figures as in the paper.

The large booth stopped more people. But of those who stopped, nearly 30% bought at the 6-flavour booth and 3% at the 24-flavour booth. The number of jams tasted was similar at both booths: on average 1.50 (24 flavours) and 1.38 (6 flavours).

The same paper has two more studies. In one, students were offered 6 or 30 topics for an extra-credit assignment. Of those with 6 topics, 74% turned in the assignment; with 30 topics, 60% did.

The authors name the study’s limits themselves. People who saw the 6-flavour booth may have believed there was something special about those six, given how many jams the store had. The 24-flavour booth may have aroused the curiosity of passers-by who weren’t interested in buying.

Barry Schwartz and The Paradox of Choice

Psychologist Barry Schwartz devotes a whole book to the topic. It is called The Paradox of Choice: Why More Is Less and appeared in 2004.

Schwartz’s thesis: we face too many options every day, and this abundance wears on our psychological well-being. The book explains this with everyday examples such as cable channels, health care plans and buying a pair of jeans.

One distinction from the book helps with interfaces too: drawing on ideas from Herbert A. Simon, Schwartz describes two types of decision makers. Those who look for the best (maximizers) weigh every possible option, and that gets harder with more options. Those who look for something good enough (satisficers) take what meets their own criteria and don’t dwell on whether something better exists.

Is the effect real?

The jam experiment took place in a single store on two Saturdays. In 2010, Benjamin Scheibehenne, Rainer Greifeneder and Peter M. Todd combined similar experiments in a meta-analysis. The study in the Journal of Consumer Research covers 63 conditions from 50 experiments and 5,036 participants.

The result: the mean effect size is virtually zero, but there is considerable variance between studies. The authors named several potentially important preconditions for the effect. They found no condition that is sufficient on its own to reliably produce it.

In 2015, Alexander Chernev, Ulf Böckenholt and Joseph Goodman published a second meta-analysis in the Journal of Consumer Psychology. They examined 99 observations and 7,202 participants. They found four factors that strengthen the effect:

• The choice set is complex.
• The decision task itself is difficult.
• The person is uncertain about what they want.
• The person wants to decide with little effort.

Taking these factors into account, the effect of the number of options became significant. The authors present this as a finding that contradicts the earlier meta-analysis.

My reading: there is no law that “many options are always bad”. When the options resemble each other and the person doesn’t know what they want, the risk rises. As Iyengar and Lepper note in their introduction, for someone looking for a particular product, more options should also increase the chance of finding it.

How it differs from Hick’s law

The two ideas are often confused. Hick’s law is about decision time: it grows logarithmically with more options. It is a relationship measured in seconds.

The paradox of choice, by contrast, is about the outcome of the decision. Does the person choose, are they satisfied with the choice, do they come back? In the jam experiment, purchases were measured.

In practice, Hick’s law describes how quickly the person chooses when the menu gets shorter. The paradox of choice describes the risk that they leave the menu without choosing anything. Both are tied to cognitive load.

Live example: three shelves, one decision

Below are three versions of the same fictional jam shelf. Choose a shelf, press Start and pick a jam for breakfast. The demo shows the time until you choose and the number of your clicks.

Three shelves, one decision

This demo works with JavaScript. If it’s off, the table on the jam experiment above shows the same idea with real figures.

A fictional jam shelf; the flavours, groups and the three “Pick” labels are made up. The clock runs from Start to “I’ll take this one”. Clicks include changing the selection, filters and confirmation; keyboard selections count too. This is your personal measurement, not an experiment and not a statistic. Nothing is stored.

There is no right answer; choose the jam you want. Try all three shelves and look at your numbers. The shelf with the filter has two shortcuts: if you know what you want, use the filter; if not, use the pick.

A note: the time of the first attempt also includes reading an unfamiliar shelf. In the second round your time will probably be shorter, because you already know the flavours. So read the numbers together with how deciding felt.

What you can do in the interface

The goal is to give the person a short path to the option they need. That usually doesn’t require removing options.

Default

Iyengar and Lepper summarise earlier studies like this: as the attractiveness of the alternatives rises, people experience conflict. They postpone the decision, search for new alternatives, choose the default or don’t choose at all.

For shipping methods or on a settings screen, preselect the option that suits most users. The default must be in the person’s interest. Quietly preselecting the expensive option would be a dark pattern.

Filter

The honest way to shrink a large catalogue is a filter. The person chooses their own criterion, and the shelf narrows accordingly. You don’t remove any product; what you don’t show right now stays one click away.

The filter itself is a choice screen too. Forty filters just move the same problem up a level. Details are in the filtering and sorting pattern.

Pick

For a person who isn’t sure what they want, a pick is the shortest path. A label such as “Good place to start” takes away the effort of comparing. Add the reason, and say “most popular” only if it is true.

Number of plans

A pricing page is a choice screen too. The more the plans resemble each other, the harder the comparison, and the decision gets postponed.

My rule: each plan speaks to a different type of user, and the difference can be said in one line. Make the recommended plan clear. Details are in the pricing page pattern.

My note

It is easy to boil the jam experiment down to “reduce the options”. Before reducing, I ask one question: does the person know what they want? If so, I give them search and filters; if not, a pick and a good default.

In my view, the real problem is often that the options can’t be told apart. Someone who sees no difference between six plans hesitates between two as well.

If you want the choice screens in your product checked from outside, the UX Audit looks at a flow from start to finish and delivers a prioritised list of fixes.