Why more options sometimes mean fewer sales

The jam display that started it

In 2000, Sheena Iyengar and Mark Lepper published "When Choice Is Demotivating: Can One Desire Too Much of a Good Thing?" in the Journal of Personality and Social Psychology. The researchers set up tasting displays offering either 24 or 6 varieties of jam and watched what shoppers did.

According to the Association for Psychological Science's account of the study, shoppers exposed to the 24-flavour display tasted more samples than those who saw only 6. But when it came to buying, the pattern flipped: shoppers were more likely to actually purchase jam when only 6 kinds were available.

The replication problem

Ten years later, Benjamin Scheibehenne, Rainer Greifeneder and Peter Todd tried to settle the question at scale. Their meta-analysis, published in the Journal of Consumer Research in 2010, pooled 50 published and unpublished experiments covering 5,036 participants.

The average effect across all of that data was, in the researchers' own words, virtually zero. Adverse consequences, they concluded, do not necessarily follow from increases in the number of options available to a chooser.

The same analysis did find a more-is-better pattern in some cases, particularly when people had clear prior preferences going in — food was named as one such category. But the authors were explicit about the limits of their own conclusion: they could not identify sufficient conditions or specific circumstances that explain when and why an increase in assortment size can be expected to reliably decrease satisfaction. In plain terms, the jam result was real in that one field setting, but it did not generalize into a dependable law.

What actually predicts the backfire

A second, larger meta-analysis in 2015 picked up where Scheibehenne's left off. Alexander Chernev, Ulf Böckenholt and Joseph Goodman, writing in the Journal of Consumer Psychology, pooled data from 57 earlier studies represented in 21 papers and asked not whether choice overload exists, but when.

Their answer was four moderating factors, reported by Kellogg Insight, the research publication of Northwestern's Kellogg School of Management, where Chernev teaches. Choice-set complexity matters more than raw count: five options each described by 50 attributes creates more overload than 50 options each described by one attribute, such as lifetime performance. Decision-task difficulty matters, with time pressure making overload worse. Preference uncertainty matters, since choosers who already know what they want compare options more easily. And decision goal matters, because gathering information feels lighter than being forced into a final trade-off.

The same research flagged a measurement problem inside the broader literature: behavioral paralysis, people refusing to choose at all, turned out to be wildly erratic across studies, making it a far less stable way to detect choice overload than tracking regret or the desire to switch to a different option afterward. Some of the inconsistency in decades of choice-overload findings may trace back to researchers measuring different things and calling them the same effect.

The number of options on offer is not, by itself, a reliable predictor of whether people buy.

What this means for a pricing page or a menu

None of the three studies gives a number of options to stay under. The number itself is not the lever. What drives the effect, according to Chernev's analysis, is how much a buyer has to process to tell the options apart, how much time pressure they feel, how clear their own preferences already are, and whether they are just browsing or expected to commit.

That reframes a common fix. Cutting a product line or a pricing menu down to three tiers does not address complexity if each tier still requires comparing a dozen attributes against the others. Kellogg Insight's reporting on Chernev's work points toward restructuring the comparison itself: helping a buyer narrow by their own stated priorities first, reducing the attributes they must weigh at once, and giving them room to decide without an artificial deadline.

The research also suggests a limit on another popular habit: counting how many visitors "bounce" from a pricing page without choosing as proof that the page has too many options. Chernev's team found that non-choice is an unreliable signal. A buyer who leaves and comes back later, or switches plans within a week, is giving a more consistent read on whether the choice set was actually too hard than one who simply didn't click anything on a single visit.

Where the evidence still runs out

The gap between the two meta-analyses matters as much as their findings. Scheibehenne's team, pooling the broadest set of existing experiments, could not state in advance which situations would produce choice overload. Chernev's team, working with a different and partly later set of studies, proposed four factors that did predict it, but a conceptual meta-analysis of existing papers is not the same as a new field experiment confirming those factors operate the way a pricing team would hope on a specific page or shelf.

What both analyses agree on is narrower than the jam study made famous: the number of options on offer is not, by itself, a reliable predictor of whether people buy. What sits around that number, how it's organized, how much time a buyer has, how sure they already are, and what they think they're being asked to do, is.