In 2000, a tasting booth at a California grocery store became one of the most-cited studies in consumer psychology. Six varieties of jam sold ten times better than twenty-four. The finding launched a thousand SaaS pricing pages and product menus. A 2010 meta-analysis of fifty studies found essentially no overall effect. A 2015 reanalysis found the effect appears under specific moderators. This is the most useful kind of replication story for anyone making product, pricing, or UX decisions --- one where the answer is neither “true” nor “false” but “it depends, and the conditions matter.”

On two consecutive Saturdays in the mid-1990s, two Stanford researchers named Sheena Iyengar and Mark Lepper set up a tasting booth at Draeger’s, an upscale grocery store in Menlo Park, California. The booth alternated between two configurations. Sometimes it displayed six varieties of jam. Sometimes it displayed twenty-four. Shoppers passing by could taste any of the displayed jams and pick up a discount coupon to buy a jar.

Then Iyengar and Lepper watched what happened. About 60 percent of passing shoppers stopped at the twenty-four-jam display. About 40 percent stopped at the six-jam display. The big display drew more traffic. But of the people who stopped at the twenty-four-jam display, only about 3 percent actually used their coupon to buy a jar. Of the people who stopped at the six-jam display, about 30 percent did.

The conversion rate at the smaller display was roughly ten times higher than at the larger display.

Iyengar and Lepper published this finding in 2000 in the Journal of Personality and Social Psychology as the central study in “When choice is demotivating: Can one desire too much of a good thing?” The interpretation was elegant. When faced with too many options, consumers got overwhelmed, struggled to compare, deferred the decision, and ultimately purchased nothing. The phenomenon was called choice overload, or sometimes the paradox of choice, and it became one of the most-cited findings in consumer psychology.

The jam study reshaped how a generation of designers, marketers, and product managers thought about product menus, pricing pages, and customer-facing decisions. Barry Schwartz’s 2004 book The Paradox of Choice extended the finding into a popular framework. SaaS companies stopped offering twelve pricing tiers and started offering three. Restaurants pruned their menus. UX designers preached “less is more.” A short, simple lab finding about jam became, for two decades, the canonical evidence that less choice is more conversion.

In 2010, a German psychologist named Benjamin Scheibehenne and two colleagues published a meta-analysis of fifty studies of choice overload. They found that across the literature as a whole, the overall effect of more choice on conversion or satisfaction was essentially zero.

This article is about what actually happens when you give people more or fewer options. Not “is choice overload real?” --- that turns out to be the wrong question. The right question is “under what specific conditions does adding options hurt conversion?” The answer is more useful than either yes or no, and it has direct implications for anyone who designs pricing pages, product menus, or customer-facing decision flows.

What the Original Study Actually Did

The founding paper is Iyengar & Lepper (2000), “When choice is demotivating: Can one desire too much of a good thing?” in the Journal of Personality and Social Psychology. The paper had three studies, not just the jam study, though the jam study became the cultural touchstone.

Study 1 (the jam study) was a field experiment at Draeger’s grocery store. The tasting display alternated between 6 jams and 24 jams every hour across two Saturdays. The dependent measures were how many shoppers stopped at the display (a measure of attraction) and how many used their tasting coupon to purchase jam (a measure of conversion). The headline finding: large displays attracted more traffic but converted far fewer customers. Approximate sample: 754 shoppers were in the store during the experimental hours; roughly 502 passed the display, of whom 249 stopped; of those who stopped, roughly 30% of small-display visitors used the coupon vs. 3% of large-display visitors.

Study 2 was a controlled laboratory experiment with college students. Participants chose between 6 or 30 essay topics for an extra-credit assignment. The smaller-set condition produced both more essay completion and higher essay quality.

Study 3 used chocolates. Participants chose 1 of 6 chocolates or 1 of 30 chocolates and rated their satisfaction. Participants in the larger-choice condition reported lower satisfaction with their selection.

Across the three studies, the pattern was consistent: large choice sets produced worse downstream outcomes (lower conversion, lower task completion, lower satisfaction) than smaller choice sets. The interpretation was that the cognitive effort of comparing many options produced decision difficulty, regret about foregone alternatives, and ultimately deferral or dissatisfaction.

This was a striking and immediately actionable finding. The jam study in particular --- with its memorable 10-to-1 conversion ratio --- became one of the most-cited consumer-psychology studies of the next two decades. Barry Schwartz’s 2004 The Paradox of Choice extended the framework into broader claims about life satisfaction and political philosophy. SaaS companies, e-commerce platforms, restaurant menus, financial products, and retirement-plan designs all incorporated “less is more” as a working principle. Choice overload became one of the most operationally useful behavioral-economics findings in business contexts.

The 2010 Meta-Analysis That Found Nothing

In 2010, Benjamin Scheibehenne, Rainer Greifeneder, and Peter Todd published “Can there ever be too much of a good thing? A meta-analytic review of choice overload,” in the Journal of Consumer Research. They aggregated 50 published and unpublished studies of choice overload, covering approximately 5,000 participants.

The result: across all 50 studies, the average effect of larger choice sets on conversion, satisfaction, or other downstream outcomes was approximately d = 0 --- essentially zero. The literature was highly heterogeneous, with some studies showing strong choice-overload effects, some showing no effect, and some showing the opposite (more choice producing better outcomes).

This was, in academic terms, a serious challenge to the choice-overload framework. If the average effect of more options was zero across a large literature, the dramatic 10-to-1 conversion ratio in the original jam study was either an outlier or the result of conditions that didn’t generalize to most contexts. The strong popular claim --- that more choice systematically hurts conversion --- was not supported by the aggregated evidence.

The Scheibehenne paper also documented substantial publication bias in the choice-overload literature. The pattern of effect sizes suggested that successful demonstrations were preferentially published and failed demonstrations were preferentially not, which would inflate the apparent effect of the construct in any narrative reading of the literature.

Defenders of choice overload pushed back. Some argued the meta-analysis included studies that weren’t really tests of the original construct. Some argued the construct had been mis-operationalized. Some pointed out that the original jam study was a field experiment with real shoppers, not a lab study with college students, and that lab studies might systematically under-detect the effect.

The field needed a more careful analysis of when the effect did and didn’t appear.

The 2015 Conditional-Effect Reanalysis

In 2015, Alexander Chernev, Ulf Böckenholt, and Joseph Goodman published “Choice overload: A conceptual review and meta-analysis,” in the Journal of Consumer Psychology. They aggregated approximately 99 effects from the broader literature and ran a moderator-analysis approach: rather than asking “is choice overload real?” they asked “under what specific conditions does it appear?”

The findings were specific and useful. Choice overload was substantially more likely to appear when four specific conditions were present:

High choice-set complexity. When the options in the set were complex, differed on multiple attributes, or were difficult to compare directly, adding more options produced more overload. Sets of options with simple attributes (a single varying dimension, easy comparison) showed minimal overload effects regardless of size.

High decision-task difficulty. When the decision required the participant to actively evaluate, weigh, and integrate information across options, larger sets produced more overload. When the decision was easy (pick a favorite, choose anything), set size mattered less.

High preference uncertainty. When participants were uncertain about their own preferences in the domain (novices, unfamiliar product categories), larger sets produced more overload. When participants had clear pre-existing preferences (experts, familiar categories), set size mattered less.

High decision goal of effort-minimization or regret-avoidance. When participants were motivated to minimize cognitive effort or avoid potential regret about foregone alternatives, larger sets produced more overload. When participants were motivated to maximize decision quality, more options could be helpful rather than harmful.

When all four moderators were present at high levels, the choice-overload effect was substantial (d ≈ 0.5 to 0.6 in Chernev’s analyses). When the moderators were absent, the effect disappeared or even reversed (more choice helped).

This was a much more useful framework than either “choice overload is real” or “choice overload is fake.” It identified the specific conditions under which adding options would hurt downstream outcomes, and the conditions under which it wouldn’t. For practitioners trying to apply the framework to real product, pricing, or UX decisions, this was the version that actually translated into action.

Why the Original Looked So Dramatic

The 10-to-1 conversion ratio in the original jam study is a famously large effect. Why did the original look so dramatic if the aggregated effect across the literature is small?

The jam study hit all four Chernev moderators at high levels. Choosing a jam at an upscale grocery is genuinely a complex decision (24 jams differ on flavor, brand, country of origin, sweetness, label aesthetics --- many dimensions). The task is moderately difficult (taste, compare, evaluate). Most shoppers have uncertain preferences about specialty jams. And the goal is effort-minimization (no one is trying to win a jam-tasting competition). The conditions of the original jam study were essentially the maximum-overload configuration. The dramatic effect size in that specific context is consistent with the moderator analysis.

The jam study was a field experiment with real consumer stakes. Lab studies of choice with hypothetical decisions or trivial rewards may under-detect real-world overload effects because the participants don’t actually have to live with their choice. The jam study captured a real-world overload condition that subsequent lab studies may have systematically under-replicated.

The sample of “stopped at display” customers was self-selected. Of the shoppers who passed the display, only the ones who stopped became part of the conversion-rate calculation. If shoppers who would otherwise have been willing to buy jam saw the 24-jam display and decided not to stop at all (because they could tell from a distance it would take too long), the conversion-rate calculation may overstate the overload effect within the population of all jam-interested shoppers. The original paper’s sample selection introduces some interpretive complexity.

Publication bias amplified the original effect’s salience. The jam study became a cultural touchstone partly because its effect was large. Smaller and null choice-overload effects in subsequent studies were less culturally amplified. The aggregated literature might look less dramatic than the famous original even if the original was conducted accurately, simply because the original was the high-end tail of the literature’s effect-size distribution.

The Honest Verdict Today

Choice overload is a real but conditional phenomenon. It is not a universal law of “more options hurt conversion.” It depends on:

  • Whether the choices are complex (multi-attribute, hard to compare)
  • Whether the decision is difficult (requires active evaluation)
  • Whether the decision-maker is uncertain about their preferences (novice, unfamiliar category)
  • Whether the decision goal is effort-minimization or regret-avoidance (rather than maximization)

When these conditions are present at high levels, adding options can substantially hurt conversion. The famous jam-study magnitudes are plausible in these conditions. When these conditions are not present, adding options has no negative effect and may have a positive effect (broader appeal, ability to match diverse preferences).

For most real product, pricing, or UX decisions, you need to ask: how complex are my options? How difficult is the decision for my user? How uncertain are they about their own preferences? What is their goal? Then design accordingly.

For SaaS pricing pages with three tiers that differ on multiple attributes and target users who don’t have strong prior preferences about feature bundles: choice overload is probably real, and a three-to-five-tier menu is probably better than ten-tier. For e-commerce category pages where the products vary on simple dimensions and users have clear preferences (size, brand, color): more options probably don’t hurt, and may help. For high-stakes decisions where users want to maximize quality (financial products, healthcare): more options can be helpful if combined with filtering and comparison tools.

The honest version of choice overload is more useful than the popular version. It gives you actual diagnostic questions to ask about your specific application.

What This Means If You’re a Strategist

Three takeaways for anyone designing product, pricing, or UX decisions.

1. Audit your specific decision context against the four moderators. Before applying “less is more” to a pricing page or product menu, run the four-question audit:

  • Are my options complex (varying on multiple attributes)? If yes, overload is more likely.
  • Is the decision difficult (requires evaluation and integration)? If yes, overload is more likely.
  • Are my customers uncertain about their preferences? If yes, overload is more likely.
  • Are my customers in effort-minimization mode? If yes, overload is more likely.

If you have four “yes” answers, the literature suggests choice overload will be a meaningful drag on conversion, and reducing options is likely to help. If you have four “no” answers, more options probably won’t hurt and may help. If you have a mix, test specifically.

2. The fix for choice overload is usually structural, not subtractive. When choice overload is present, the popular advice is “remove options.” This is often suboptimal. The better fix is usually to keep options but reduce the decision complexity. Filtering tools. Sorted-by-fit recommendations. Default selections for typical users. Comparison tables that highlight relevant differences. Decision-aid widgets that ask the user a few questions and then narrow the recommendation. These structural fixes preserve the value of having options (broader appeal, ability to match diverse preferences) while reducing the cognitive load on each individual user.

A 25-product category page with good filtering will often outperform a 5-product category page with no filtering, because the filtering reduces effective choice-set complexity to whatever the user actually needs to evaluate. This is a more powerful intervention than simply cutting product variety.

3. Test rather than assume. The choice-overload literature is the single best case for why behavioral-economics applications should be empirically tested in your specific context rather than assumed. The original jam study findings cannot tell you what will happen to your SaaS pricing-page conversion when you change from five tiers to three. The effect size, direction, and even sign of “less is more” is highly context-dependent.

For any decision where set-size matters (pricing pages, feature menus, product categories, signup flow steps), the right move is to A/B test rather than apply a heuristic from the literature. The literature gives you good hypotheses about which direction to test in (smaller is probably better when the four moderators are high, larger may be better when they’re low) but the magnitude --- and sometimes the sign --- has to be measured in your context.

The choice-overload story is the cleanest available demonstration of why behavioral economics is best treated as a source of hypotheses to test rather than as a set of laws to apply.

Sources

This article is part of an ongoing series on famous behavioral-science studies. Other entries cover power posing, the marshmallow test, ego depletion, the facial feedback hypothesis, Bargh elderly priming, growth mindset, loss aversion, the Stanford Prison Experiment, the bystander effect, and the Mozart Effect. The full hub lives at /replication-crisis/.

If you’ve designed pricing, menu, or product-selection decisions on the assumption that less choice is always better and want a careful audit of whether the choice-overload conditions actually apply to your context, book a consultation.

FAQ

So is choice overload “real” or not? Real, but conditional. The phenomenon appears under specific conditions (complex options, difficult decisions, uncertain preferences, effort-minimization goals). When those conditions are absent, more options don’t hurt. The universal “less is more” framing is not supported.

Should I still use the famous jam study in talks and writing? With caveats. The jam study is real and the original methodology was reasonable. The issue is that the effect size in the original study was at the high end of what the aggregated literature shows, because the conditions of the original were essentially the maximum-overload configuration. If you cite the study, cite the Chernev 2015 moderator analysis alongside it so your audience understands that the dramatic 10-to-1 ratio depends on specific conditions.

What’s the optimal number of choices for a pricing page? There is no universal answer. The literature would suggest three to five tiers for typical SaaS pricing --- complex enough to differentiate offerings, simple enough to avoid overload. But the right answer for your specific product depends on how complex your tiers are, how clear your target customers are about what they need, and how much filtering/guidance you provide. Test rather than guess.

Has Sheena Iyengar responded to the meta-analyses? Yes. Iyengar has acknowledged the moderator analysis and has incorporated it into her subsequent work. She has not retracted the original study and her position is similar to Mrkva’s defense of loss aversion: the phenomenon is real but more conditional than the popular versions imply. Her 2010 book The Art of Choosing engages with the more nuanced picture.

What about Barry Schwartz’s The Paradox of Choice? Schwartz’s book extended choice overload into a broader cultural argument about modern life. The empirical claims about specific consumer-decision contexts are subject to the same caveats as the underlying literature. The broader cultural argument is more philosophical than empirical and is harder to test directly. Treat The Paradox of Choice as an interesting framework rather than as established science.

Does this mean I should add more options to my product? Not necessarily. It means you should think about your specific decision context. If your options are simple (single-attribute differences), your users are experts in the category, and they’re trying to make a high-quality decision, adding options probably won’t hurt and may help. If your options are complex (multi-attribute), your users are novices, and they’re trying to minimize effort, fewer options or better filtering will help.

replication-crisis behavioral-economics consumer-psychology pricing-strategy ux-design

Free Tool

Built for Experimentation Teams

GrowthLayer is the experimentation platform I built for CRO teams --- test management, AI-powered insights, and pattern recognition across your entire program.

Explore GrowthLayer → (opens in new tab)

· Start Free →

Share this article

LinkedIn (opens in new tab) X / Twitter (opens in new tab)

Copy link

Go deeper

Methodology The PRISM Method Case Studies $30M+ in Results Work Together Services & Mentoring

Experimentation and growth leader. Builds AI-powered tools, runs conversion programs, and writes about economics, behavioral science, and shipping faster.

About LinkedIn Newsletter

← Previous

Ego Depletion: How Willpower Became a Glucose Tank --- And Why That Story Collapsed

Next →

The Bystander Effect and Kitty Genovese: How One Mistaken Newspaper Story Created a Field --- and What CCTV Footage Reveals Now

replication-crisis behavioral-economics consumer-psychology pricing-strategy ux-design
Share this article
LinkedIn (opens in new tab) X / Twitter (opens in new tab)
Atticus Li

Experimentation and growth leader. CXL-certified CRO practitioner, Mindworx-certified behavioral economist (1 of ~1,000 worldwide). 200+ A/B tests across energy, SaaS, fintech, e-commerce, and marketplace verticals.

About LinkedIn Newsletter