A product list decides what shoppers can compare without opening another page. Rows may make price and features easier to line up. Visual grids or horizontal assortments may help someone inspect the range. The useful format depends on the job.
For an operator, the decision is which layout helps customers choose suitable products while meeting the store’s commercial goals. Speed, cheaper selections, variety and satisfaction are separate outcomes. A single conversion number cannot explain all of them.
🔎 What research says
Match the layout to the comparison task. Test aligned rows for feature and price comparisons, and horizontal displays when customers want several varieties. Measure choice satisfaction and commercial outcomes separately; the evidence does not establish one format that wins for every task or device.
🗺️ In this guide
1. Align the attributes shoppers need to compare
📈 Recommendation
Test a row-based list when shoppers need to compare price and specific features. Keep the decisive attributes visible in the same place for each product.
Illustrative store test: a headphone seller could compare its existing tiles with rows showing price, battery life and connection type. Keep the assortment, prices and product information the same.
Check whether shoppers can correctly identify an option meeting a stated need, how long that takes, and which products they choose. Track paid orders and margin separately; cheaper customer choices do not automatically improve seller economics.
🎓 Findings
Schmutz and colleagues (2010) compared list and matrix presentation in an online shopping study, followed by a laboratory eye-tracking study. Their primary institutional abstract reports lower mental workload and more economical product selections with lists. It suggests that lists encouraged comparison.
The accessible primary record does not provide a verified size of the price or workload difference. It cannot tell a seller how much cheaper the selected products were, how much faster shoppers worked, or how paid revenue changed.
🧠 Why it works
Our interpretation is that consistent attribute positions remove some of the work of finding the same fact repeatedly. For headphones, that means comparing battery life rather than hunting for it. The proposed column arrangement is an application, not the exact interface verified in the accessible abstract.
✋ Limitations
Full methods and tables were not recovered after checking the publisher, author searches and the University of Basel repository. Payment conditions, detailed task, exact estimates and uncertainty remain unresolved. This supports a conditional comparison test, not a universal list-over-grid rule or a promised sales gain.
2. Test horizontal displays when shoppers want a mix
📈 Recommendation
Test a horizontal assortment when customers commonly choose several varieties. Keep a separate test for customers selecting one familiar favorite.
Illustrative store application: a snack seller could compare two layouts for building a mixed box, while keeping products, prices and the permitted box size unchanged. Count distinct flavors and total units separately.
Check the actual phone layout. A horizontal strip that conceals most options behind an unnoticed swipe is a different treatment from an assortment visible at once. Do not assume a desktop or physical-display result transfers.
🎓 Findings
Deng and colleagues (2016) studied display direction. In their laboratory mini-store, 271 students received $2, paid 25 cents per candy and kept unspent money. For people allowed several candy types, horizontal presentation produced 3.00 distinct types versus 2.36 vertically: 0.64 more types per participant. The difference was statistically clear; total quantity showed no clear difference across the four study conditions.
A separate store survey found 4.44 sanitizer units bought by horizontal-display shoppers versus 2.75 by vertical-display shoppers, a 1.69-unit gap. Those 67 customers chose which display to use. This is an association among surveyed buyers, not proof that changing display direction caused extra sales.
🧠 Why it works
Making several alternatives easy to inspect may help someone assemble a varied pack. That is a different objective from finding one known favorite. More distinct flavors can also fit inside the same number of units; variety and basket size are different business results.
✋ Limitations
The mini-store used familiar candy and a small experimental budget; the survey involved one physical retailer. Neither estimates the effect of an online carousel. The paper treats mobile transfer as an open question. The quantity null leaves uncertainty and does not prove identical quantities.
3. Measure satisfaction alongside selection and sales
📈 Recommendation
Choose a layout using both decision quality and commercial results. Ask whether shoppers are happy with their choice, then check orders, cancellations and returns separately.
Illustrative store test: after a headphone comparison, ask a short choice-satisfaction question. Separately follow the order through cancellation and return; avoid treating a favorable survey response as a completed sale.
For a rollout decision, inspect results by product task and device. A lower-priced choice, a higher satisfaction score and more profitable retained orders can move in different directions. Decide which outcomes and tradeoffs matter before choosing a winner.
🎓 Findings
In the same Deng study, multi-type choosers rated satisfaction 7.92 horizontally versus 7.37 vertically on a 1–9 scale: a 0.55-point gap, not a purchase-rate increase. For one-type choosers, the display difference was not statistically clear. This is another outcome from the same study, not an independent replication.
Jia, Fang and Ouyang (2025 issue; online 2024) report higher choice satisfaction with horizontal displays of comparable products across five experiments. A cue marking the end of the display weakened that effect. Their accessible appendix documents satisfaction questions and separate behavioral measures, but does not supply the group effects needed to quantify a sales or cancellation benefit.
🧠 Why it works
A shopper can settle on an option without choosing the cheapest one. Satisfaction measures that sense of resolution; an order record measures a different event. Keeping both makes a useful interface improvement less likely to be confused with a commercial gain.
✋ Limitations
The newer paper’s main full text was not recovered; its primary abstract and e-companion were reviewed. Exact group effects, uncertainty and complete payment procedures remain unresolved. Its satisfaction result and the earlier list study’s economy result use different designs and outcomes, so they do not establish a contradiction or a universal winning layout.
💡 The takeaway
Choose the comparison task before choosing the layout.
Run a controlled comparison with the same eligible products and information. Inspect what shoppers select, how confidently they decide and what happens to paid orders. Keep mixed-pack and single-favorite tasks, and phone and desktop experiences, visible in the results.
📚 Sources and study types
Schmutz, P., Roth, S. P., Seckler, M., & Opwis, K. (2010). Designing product listing pages—Effects on sales and users’ cognitive workload. International Journal of Human-Computer Studies, 68(7), 423–431. Online shopping comparison and laboratory eye tracking; primary University of Basel record supports direction only; complete estimates not recovered.
Deng, X., Kahn, B. E., Unnava, H. R., & Lee, H. (2016). A “Wide” Variety: Effects of Horizontal versus Vertical Display on Assortment Processing, Perceived Variety, and Choice. Journal of Marketing Research, 53(5), 682–698. Five field and laboratory studies. Numerical results checked in the author-hosted primary manuscript, Study 1 and Study 5. The store-display survey did not randomly assign buyers to a display.
Jia, Y., Fang, Y., & Ouyang, J. (2025; first online 2024). How Product Display Orientation Affects Customers’ Choice Satisfaction in Online Purchase: A Choice Closure Perspective. Information Systems Research, 36(3), 1587–1611. Five experiments including eye tracking; primary abstract and e-companion reviewed; main estimates remain unavailable.
