Three ways of letting a shopper pick a size. Same products, same prices, same retailer, same mobile screen. The only thing that changed was whether the range was visible, hidden, or revealed on demand, and it changed how people shopped so completely that each layout produced a different kind of shopper.
Most variant-selector debates are argued on taste and screen real estate. We ran them as an experiment instead, on a US mass retailer's mobile site across three categories, and the behavioural data separated the three layouts into three distinct shopping modes. Naming them turned out to be more useful than ranking them.
Hidden variety: the task terminal
In the first layout, tapping a product opened a modal containing the full range, then sent the shopper to their cart once they added something. Variety exists, but only after you commit to looking for it.
This produced the fastest, most transactional trip in the study: 94 seconds against 109 in the current experience, and the highest penetration, 61% against 53%. It also produced the smallest basket, 1.7 products against 2.1, and the lowest spend, $21.10 against $27.30.
The reason is what the shopper meets first. The modal opens on a pre-selected option, and that option is often the large pack, so the first number a shopper sees is the highest price in the range. We started calling the reaction price panic: the shopper's opening impression of the product is its most expensive form, and the exit to the cart is one tap away. The layout is efficient at finishing a purchase and hostile to starting another one.
Persistent variety: the interactive shelf
In the second layout, the alternatives sat on the product card itself as small text pills, and tapping one kept the shopper on the results page.
This kept the basket intact: 1.9 products and $26.80 spend, both essentially level with the current experience, and category value came in at just 1% below it. Shoppers toggled between options while staying on the shelf, comparing the small against the large without leaving the aisle. Where the modal triggered price panic, visible pills produced something closer to value validation, the shopper satisfying themselves that the option they picked was the sensible one.
It also earned the best experience scores in the study. Shoppers found it easier to navigate and easier to pick a size, which is why we recommended it as the basis for further testing in one category even though it produced no incremental value.
Structured variety: the detailed menu
The third layout revealed the range as a vertical drop-down list, again without leaving the results page.
This was the slowest route through the store, 107 seconds, and the most deliberate. It held basket size at 2.1 products, matching the current experience exactly, on penetration of 53%. A clear list invites reading rather than reflex; shoppers behaved analytically, and the category ended 3% down in value, the mildest loss of the three.
| Layout | Trip time | Penetration | Basket | Category value |
| Hidden variety (modal) | 94 seconds | 61% | 1.7 products | 11% lower |
| Persistent variety (pills) | — | — | 1.9 products | 1% lower |
| Structured variety (drop-down) | 107 seconds | 53% | 2.1 products | 3% lower |
| Current experience | 109 seconds | 53% | 2.1 products | — |
The pattern worth remembering
Line the three up and the trade is obvious. The layout that converted best produced the smallest basket. The layout that took longest protected the basket best. Speed and value pulled in opposite directions across every category we tested, which is an uncomfortable finding for anyone whose UX brief is written purely in terms of friction reduction.
There is a second pattern in where the sales came from. In the modal layout, 44% of category buyers bought from inside the modal and 28% bought from nowhere else. With pills or a drop-down, that inverted: around a third bought through the selector and roughly a third still bought a single product straight off the shelf. The selector is not a neutral convenience. It becomes the shelf, and whatever it happens to show first becomes the default.
The cost nobody budgets for
Every one of these layouts does the same structural thing: it collapses individual product tiles into a single card. Total facings fell 53% in laundry and 73% in nappies.
That cost is not shared equally. The largest detergent brand went from 91 tiles to 9 plus a selector, a 62% reduction. A mid-sized brand lost 27%. Predictably, the big brands lost value in every tested layout while several small ones gained. Consolidating a shelf is a redistribution of visibility, and visibility is what large brands have been paying for all along.
How to use this
Decide what the layout is for before you choose one. If the goal is completing a difficult purchase quickly, hidden variety does that better than anything else we tested, and you should expect to pay for it in basket size. If the goal is protecting basket value while making a wide range navigable, keep the alternatives visible on the card and keep the shopper on the shelf.
Then check two things the layout choice will not tell you: how many facings you just removed, and which option the selector opens on. In this study, that second detail, the pre-selected large pack, was doing more damage than the layout itself.
None of this came out of asking shoppers which selector they preferred. It came from watching what each one made them do.