How Choice Overload on Category Pages Kills Conversions (And What to A/B Test)
Quick answer
Choice overload on ecommerce category pages — triggered by large product grids, unhelpful default sorting, and hidden filters — causes decision paralysis that pushes visitors to leave without buying. Five A/B tests can measurably reduce this friction: reducing the initial product count shown, changing the default sort order to a social-proof signal, introducing editorial "staff pick" callouts, making filters more prominent, and increasing product card information density.
Key takeaways
- The famous 2000 jam study by Iyengar and Lepper demonstrated that offering 24 options produced a purchase rate of 3%, while 6 options produced 30% — a 10x difference driven by reduced cognitive load, not reduced appeal.
- Category pages are where choice overload most commonly manifests in ecommerce, because they ask visitors to self-navigate hundreds of products with minimal editorial guidance.
- Structured A/B testing is the only reliable way to find which reduction strategy resonates with your specific audience — there is no universal "right" number of products to display.
The phrase "too much choice" gets used loosely in conversion rate optimisation discussions, but the underlying psychology is well-established and the ecommerce implications are significant. In their landmark 2000 study published in the Journal of Personality and Social Psychology, Sheena Iyengar and Mark Lepper set up tasting booths at a gourmet market, alternating between displays of 24 varieties of jam and 6 varieties.
The larger display attracted more initial interest — 60% of passersby stopped, compared to 40% at the smaller display. But when it came to actually purchasing, the numbers reversed dramatically: 30% of visitors to the 6-jam booth made a purchase, compared to just 3% of those who encountered the 24-jam booth.
More options produced more browsing and far less buying.
Barry Schwartz formalised the broader theory in his 2004 book The Paradox of Choice: Why More Is Less, arguing that beyond a certain threshold, additional options impose cognitive costs — the effort of evaluating, comparing, and anticipating regret — that outweigh the practical benefits of having more to choose from. The result is not satisfaction but anxiety, and the easiest resolution to anxiety is to defer the decision entirely.
In a physical store, deferral means putting the item back and walking away. On a website, it means closing the tab.
Ecommerce category pages are the environment where this phenomenon plays out most visibly. A visitor landing on a clothing brand's "Women's Jackets" page, or a homeware retailer's "Dining Tables" category, is immediately confronted with a grid of products — often forty, sixty, or more — with minimal signal about where to start.
Unlike a physical store where spatial arrangement, staff recommendations, and inventory constraints naturally curate the experience, a digital category page is almost infinitely expandable. The cost of adding another product to the grid is near zero, so the grid grows, and the cognitive load grows with it.
The good news is that this is one of the more tractable conversion problems in ecommerce. Choice overload is not caused by your products being unappealing — it is caused by presentation.
Changing how products are surfaced, sorted, and filtered costs little to implement and lends itself naturally to A/B testing.
How choice overload shows up on ecommerce category pages
The most obvious manifestation is the product count itself. A grid of 48 products with no editorial hierarchy looks, to a new visitor, like a research project.
Every card demands a small amount of attention — a glance at the image, a reading of the name, a note of the price — and the cumulative attentional cost of scanning the page is high before any decision is even approached. Visitors who do not have a clear existing preference (which is most of them) have no anchor, no starting point, and no reason to prefer one product over another beyond what the grid happens to surface first.
Default sort order compounds the problem. Many category pages default to "Newest first" or "A–Z" sorting, which serves internal inventory logic rather than the visitor's decision-making needs.
Alphabetical sorting offers no signal about quality, popularity, or value. Newest-first privileges recent additions regardless of whether they are the best options for an uncertain buyer.
When the default sort order carries no social proof or editorial judgment, the visitor receives no implicit recommendation — they are left to construct their own preference from scratch.
Filters are a natural antidote to large product sets, but on many sites they are either hidden behind a "Filter" button that requires an interaction to reveal, or presented as a collapsed accordion at the page's side. Baymard Institute's research on product list pages consistently identifies poor filter usability as one of the top reasons shoppers abandon category pages without converting.
When a visitor cannot easily narrow the field, the field remains large, and the cognitive load remains high.
What to A/B test to reduce choice overload on category pages
Test 1: Reduce the initial product count visible
The principle behind this test is straightforward: present fewer options to reduce the initial cognitive load, and trust that visitors who want to see more will paginate or scroll. Research on choice overload, including Iyengar and Lepper's foundational work, consistently shows that smaller choice sets produce higher rates of commitment, even when the broader assortment remains theoretically accessible.
In practice, this means comparing a control showing 40–50 products on initial page load against a variant showing 12–16 products — either through stricter pagination or by deferring additional items behind a "Load more" interaction on infinite-scroll pages. The variant reduces visual overwhelm at first contact while preserving access to the full catalogue for motivated browsers.
Pairing the reduced count with a clear product total ("Showing 16 of 142 results") reassures visitors that the full range is available on demand.
The test: Control shows 40–50 products on initial page load; variant shows 12–16 products with clear pagination or a "Load more" trigger below.
Your hypothesis is that reducing the initial product count lowers decision paralysis and increases the rate at which visitors proceed to a product detail page, leading to a higher overall category-to-product click-through rate and a measurable improvement in session-to-purchase conversion.
See it in practice → The relationship between product volume and conversion is explored further in Mida's guide to comparison fatigue and cart abandonment.
Test 2: Default sort order — social proof vs. recency or alphabetical
Sort order is one of the most overlooked variables on a category page, partly because it is easy to dismiss as a neutral infrastructure decision rather than a conversion lever. But the default sort is the editorial decision a site makes for visitors who have not expressed a preference — and that describes most visitors, most of the time.
Defaulting to "Bestsellers" or "Most reviewed" changes the implicit message the page sends. Rather than presenting products in an order that communicates nothing about quality or popularity, the page is now telling the visitor: here are the products that other people have already evaluated and chosen.
This is a form of social proof applied at the category level, and it gives uncertain visitors a meaningful starting point without requiring any action on their part.
The effect is particularly pronounced for categories where visitors feel underqualified to evaluate options independently — consumer electronics, nutritional supplements, baby products — where popularity functions as a proxy for quality. For fashion and lifestyle categories where personal taste matters more than consensus, "Most reviewed" may be a better default than "Bestsellers", since it signals community engagement without implying a single correct choice.
The test: Control uses "Newest first" or alphabetical default sort; variant defaults to "Bestsellers" or "Most reviewed."
Your hypothesis is that a social-proof-based default sort reduces the effort required to identify candidate products, lowering the time-to-first-product-detail-page-view and increasing purchase conversion among visitors who do not manually change the sort order.
Test 3: "Staff pick" or editorial callout
One of the most direct ways to counteract choice overload is to make an explicit editorial recommendation. Many physical retailers do this naturally — a "Staff favourite" card on a shelf, an endcap display, a "New & Notable" section — but ecommerce category pages rarely replicate the practice at scale.
This test introduces a small curated section at the top of the category page: two or three products highlighted with a badge ("Our pick", "Best value", "Top rated") and, optionally, a brief line of editorial copy explaining why they were selected. The goal is not to replace the broader grid but to give visitors who feel overwhelmed an immediate answer to the question "where should I start?" without forcing them to evaluate the full assortment first.
From a cognitive load perspective, the editorial callout performs the same function as a knowledgeable shop assistant who says "most people start with this one."
The badge label matters. "Our pick" signals brand confidence.
"Best value" appeals to price-sensitive visitors. "Top rated" leverages the same social proof mechanism as a bestseller sort.
Testing two badge variants in a follow-up experiment can help identify which framing resonates most with your audience.
The test: Control shows a standard product grid with no editorial highlight; variant surfaces 2–3 highlighted products with a badge label at the top of the grid before the main product list begins.
Your hypothesis is that editorial callouts reduce the proportion of visitors who leave the category without viewing any product detail page, by providing an immediate, low-effort starting point for uncertain shoppers.
See it in practice → For an introduction to identifying high-impact tests for your category pages, see Mida's guide to CRO research.
Test 4: Filter prominence — persistent sidebar vs. collapsed/hidden
Filters are the primary tool visitors can use to reduce the choice set themselves, but their effectiveness depends entirely on whether visitors use them — and usage is strongly correlated with discoverability. A filter panel hidden behind a button or collapsed by default introduces an extra step that a significant proportion of visitors will not take, particularly on mobile.
The core question this test addresses is whether making filters more immediately visible and accessible — either as a persistent sidebar on desktop, or as an above-the-fold horizontal strip showing the most commonly used filter dimensions (size, colour, price) — increases filter engagement and, downstream, conversion. Baymard's product list research supports the principle that above-the-fold filter presentation increases the proportion of visitors who engage with filtering, which in turn reduces the effective choice set that each visitor is working with.
The implementation varies by device. On desktop, a persistent left-hand sidebar with common filter facets already visible (not collapsed) is the most common high-performing pattern.
On mobile, a sticky horizontal strip with the two or three highest-usage filter types — often size and price range — performs better than a full sidebar, which takes up too much screen real estate.
The test: Control uses collapsed or hidden filters requiring an explicit interaction to reveal; variant presents key filters (size, colour, price range) as a persistent sidebar on desktop or an above-the-fold strip on mobile, visible without any interaction.
Your hypothesis is that making filters immediately visible increases filter engagement rates, reduces the average product count that visitors are working through, and improves the rate at which filtered sessions result in a product detail page view.
Test 5: Product card information density
Product cards in a category grid make a series of micro-decisions: which attributes to show, how prominently to display the price, whether to include a review count or star rating, whether to surface any secondary information like "Free shipping" or a key spec. Minimal cards — image plus name plus price — are clean but they push all evaluation to the product detail page, requiring an additional click to answer basic questions.
Richer cards that include a review count, star rating, price, and one or two key attributes allow visitors to do more preliminary filtering directly on the category page, without navigating away. For categories where certain attributes are decisive — wattage for electronics, material for outdoor furniture, thread count for bedding — surfacing those attributes on the card can significantly reduce the number of clicks required to identify candidate products.
The trade-off is visual density: richer cards take up more space and can themselves become overwhelming if overdone. The goal is to surface only the attributes that are genuinely decision-relevant for that specific category.
The test: Control shows minimal product cards (image, name, price); variant shows enriched cards including review count, star rating, price prominence, and 1–2 key specs relevant to the category.
Your hypothesis is that higher-density product cards reduce the number of detail-page visits required before a visitor identifies a product to purchase, improving the ratio of product detail page visits to add-to-cart events.
Free A/B Testing Tool
Run your next A/B test the right way
Visual editor, 15 KB script, GA4-native — and free forever up to 100,000 monthly visitors. No developer required.
Running these tests
These five tests address choice overload at different points in the category page experience — the initial product count, the implicit recommendation embedded in sort order, the presence of editorial guidance, the accessibility of self-directed filtering, and the information available per product. They are not mutually exclusive, but testing them simultaneously would make it difficult to attribute results cleanly.
A reasonable sequencing approach is to start with the test that addresses the most severe friction point in your current analytics.
If your session recordings or heatmaps show visitors scrolling past the full product grid without clicking anything, reducing the initial product count or introducing staff picks is likely the highest-priority test. If filter engagement is low in your click data, filter prominence is the more direct lever.
If visitors are clicking through to many product detail pages without converting, product card information density may be reducing the quality of the consideration set rather than the quantity of effort required.
Most of these tests require only front-end changes — CSS adjustments, sort-order configuration, badge overlays — which means they are among the faster category page experiments to implement. Running each test for at least two full weeks, and segmenting results by device type, will give you the most interpretable results.
For a structured approach to deciding which test to run first, Mida's guide to CRO research covers how to identify the highest-impact pages and friction points in your funnel before committing to a test roadmap. Additional category page test ideas are collected in the A/B Testing Idea Bank.
Ready to test? Try Mida free — now available even if you don't have an account. Or browse the A/B Testing Idea Bank.
FAQs
Q: Is choice overload a real phenomenon or just a theory?A: The foundational research is well-established. Iyengar and Lepper's 2000 jam study, published in the Journal of Personality and Social Psychology, produced one of the most replicated findings in consumer psychology: a smaller choice set (6 options) produced a purchase rate ten times higher than a larger set (24 options). Barry Schwartz's 2004 synthesis in The Paradox of Choice extended the theory across domains from consumer goods to healthcare. A 2015 meta-analysis by Chernev, Böckenholt, and Goodman across 99 studies confirmed a robust overall choice overload effect, while also noting that the effect size varies by context, product category, and how difficult preferences are to articulate. The directional finding — that beyond a certain threshold, more options reduce the rate of commitment — is well supported.
Q: Does reducing product count mean losing sales to competitors who show more?A: Not necessarily. The concern is that visitors will assume you have a limited assortment, but this depends on implementation. Clear pagination, accurate product counts ("Showing 16 of 142 results"), and visible filter options all signal that a full assortment exists. The goal is not to hide products but to reduce the cognitive cost of the initial page encounter. Visitors who want to browse more can do so; the default experience is simply less overwhelming. The jam study outcome is instructive here: the smaller booth still offered 6 choices. Reduction is about managing the initial set size, not eliminating breadth.
Q: Which product categories are most affected by choice overload?A: Categories with high attribute overlap — where many products look similar and decision-relevant differences are not immediately visible — tend to be most affected. Fashion, homewares, and consumer electronics are commonly cited examples. Categories where visitors arrive with strong prior preferences (searching for a specific product by name, for example) are less susceptible, because the visitor has already done substantial evaluation work before landing on the page. If your analytics show a high bounce rate from category pages combined with low filter engagement, choice overload is a likely contributing factor.
Q: How long should I run these tests before drawing conclusions?A: Standard A/B testing best practice is to run a test until you have reached statistical significance (typically 95% confidence) and a sufficient sample size to detect the minimum effect size you care about. For category pages with moderate traffic, this often means two to four weeks. Running tests for less than a full business cycle introduces day-of-week and promotional-period bias. Stopping early because a result looks promising introduces significance-chasing bias. Mida handles significance calculation automatically, flagging when a test has reached reliable conclusions.
Q: Can I run more than one of these tests at the same time?A: You can, using a multivariate or multi-page test design, but interpreting the results becomes more complex. If two variables interact — for example, if richer product cards only improve conversion when filters are also prominent — a simple A/B split on each variable independently will not detect the interaction. If your traffic supports it, a 2x2 factorial design can isolate individual effects and detect interactions simultaneously. For most ecommerce sites, sequential testing of the highest-priority variables is the more straightforward starting point.
Sources
- Iyengar, S. S., & Lepper, M. R. (2000). When choice is demotivating: Can one desire too much of a good thing? Journal of Personality and Social Psychology, 79(6), 995–1006. doi.org/10.1037/0022-3514.79.6.995
- Schwartz, B. (2004). The Paradox of Choice: Why More Is Less. HarperCollins.
- Chernev, A., Böckenholt, U., & Goodman, J. (2015). Choice overload: A conceptual review and meta-analysis. Journal of Consumer Psychology, 25(2), 333–358. doi.org/10.1016/j.jcps.2014.08.002
- Baymard Institute. Product List & Filtering research collection. baymard.com/blog/collections/product-list
