Who decided the order your products appear in? It was not you

In short

• The category page is your real shop window: it is where acquisition traffic lands and browses, far more than the home page ever does.
• Its sort order is almost always a default setting: date added, alphabetical, or a “relevance” that nobody can explain.
• The top positions take the bulk of the clicks, exactly as they do on a search results page. What sits at the top decides what sells.
• Physical retail has measured why: where every location is equally accessible, it is where the products sit, not the architecture of the place, that drives buying behaviour.
• It is the fastest lever before the fourth quarter: reordering needs no new content, no budget and no redesign, just a rule someone has actually chosen.

Ask the question in your next meeting: in what order do our products appear, and why? The silence that follows is where this article starts. In a physical shop, nobody would let chance decide what goes in the window and what goes in the stockroom. Online, the platform’s out-of-the-box configuration has been deciding it since day one.

What the default sort order is really doing

Every default has its own pathology, and you can spot each one in a single visit.

What physical retail already knows, and what does not carry across

Shop merchandisers have been arranging windows and walkways for a century. Three measurements tell you which of their habits survive the move to a screen, and a fourth tells you how the visitor feels about it.

First, when every location is equally good, placement becomes the only lever. A team of British researchers, including Alan Penn of University College London, followed 481 supermarket shoppers on the store’s own cameras, from the door to the till, then interviewed them on the way out. They correlated the building’s layout with the movement that actually happened. The link is weak for circulation and all but non-existent for product interactions, quite simply because every useful aisle offers the same level of accessibility. Their conclusion fits in a sentence: there are two systems, one for moving around and one for choosing, and it is product placement that drives the second.

That is precisely the situation of a category page. A grid in which every cell is technically equivalent, and where the order is therefore the only lever left.

Second, making people travel further makes them buy more, and this is not a merchandiser’s hunch. An American team recorded real shopping paths by radio-frequency tracking, then ran a field experiment: a coupon that forced shoppers off their planned route produced $21.29 in unplanned spending, against $13.83 for a coupon sitting on the route. The same work notes that a shopping trip covers less than half of the store’s areas.

That is the measured version of the move every visual merchandiser knows: put eye-catching products at the back of the space so that customers cross the whole range on their way to them.

Third, what does not carry across is the constraint. A shop imposes the route and never has to negotiate it: the customer came, they parked, they will walk. A page imposes nothing, the exit is one click away, and forcing is paid for on the spot. Research on forced exposure online shows that it triggers psychological reactance, meaning a reflex to take back control, and then avoidance. Other work measures how irritation produced by a site’s design directly degrades satisfaction, trust and purchase intention. Scrolling is not walking.

What does carry across is the principle that makes the back-of-the-store trick work in the first place: contrast. A product rises because it breaks with its neighbours, not because it is beautiful.

How the visitor feels while this happens

The supermarket study asked every shopper about their satisfaction on the way out, and the answers split by type of trip. They are counter-intuitive.

Two further studies explain why the same staging pleases one shopper and annoys another. A stimulating environment is pleasant for someone who came to browse and unpleasant for someone who came to complete a task; and the mechanism behind that unpleasantness is a loss of perceived control, which leads on to stress, less pleasure and less intention to return.

Translated for your category page: the visitor who arrived from a precise search and the visitor browsing from section to section are not the same customer, and your sort order speaks to both at once. Hence the simple rule: arrange the top positions, never lock the path. A visible sort, filter and search are what make an arranged order bearable for the first of the two.

One last fact usefully contradicts the caricature of the forced route. The retailer most associated with the compulsory path opened city-centre stores without one, where shoppers went straight to what they wanted, then closed those in Madrid, Shanghai and Warsaw: its own interviews and surveys said customers wanted to be guided. Guiding is not constraining, and guiding is exactly what an arranged order does.

The trap of the beautiful arrangement

One practice is spreading, and it looks like progress: arranging the category by colour harmony, running together products that share a palette, producing a page that holds together. The result is often genuinely beautiful.

It is also the opposite of what physical retail does. The shop merchandiser uses colour to break the line and create a hook that pulls the eye further along. Arranging by tonality does the reverse: it smooths, and a smoothed screen has no hook left. The descent flattens out.

That does not make it a mistake. It is a trade-off, and a legitimate one if the stated objective is brand image. It stops being legitimate if the stated objective is revenue, and the way to settle it is not a matter of taste. Three checks, all of them in your own analytics: the click rate of positions 1 to 6 against the ones below, the average scroll depth on the category, and the share of revenue produced by the first six slots. If the arrangement does not make people go down, it is only decoration.

Two worked examples

The two situations below are common patterns, described to show the reasoning. They are not client results, so there are no performance figures to copy: what transfers is the way the call is made.

Example 1: a “Dresses” category arranged by palette, late September

The order found. Twelve dresses running through a gradient of beiges and matching tones, very good-looking in a screenshot. In positions 1 and 2, two summer styles at the end of their run, two sizes left in each. The wool dress that has just arrived sits in position 7.

What is wrong with it. The page is uniform, so the eye has no hook and stops around the sixth tile. The in-season product, the one carrying the margin and the stock, sits just below that waterline. And the first two positions are occupied by remnant stock: every sale there breaks the next size, so the best window on the site is working to destroy its own range.

The order set. The wool dress in position 1. A strong-coloured style in position 3, to break the gradient and send the eye downward again. The two best sellers of the last thirty days in positions 5 and 6. End-of-run styles moved past position 12 without being hidden, because they still sell to whoever is looking for them. The page is less beautiful than it was. It makes people go down.

Example 2: a “Running shoes” category left on price, low to high

The order found. The first six models are the six cheapest in the catalogue, two of them entry-level lines being cleared.

What is wrong with it. The visitor reads a price ladder before reading an offer, and the brand positions itself at the bottom of the market on its most visited page, entirely on its own. Sorting by price is not the problem: it is an excellent filter, which visitors switch on when they need it. It turns harmful when it doubles as the default order, that is, when it is imposed on someone who never asked for it.

The order set. Open on the best-selling mid-range model. Deliberately alternate price levels across the first six positions, to break the staircase reading. And one guardrail specific to footwear: no model with half its sizes unavailable may sit in the first row, because a first row that leads to missing sizes costs more than an empty slot.

In both cases the verification is the same, and it is done before and after: average scroll depth on the category, and the share of revenue produced by the first six positions. That is what separates an arrangement that sells from an arrangement that decorates.

Write the rule first, choose the tool second

A good category order carries out a policy, it does not invent one. That policy comes down to three questions.

What are we pushing? Margin, newness, clearance, season: the answer changes through the year, and that is perfectly normal. Going into the fourth quarter, the answer is usually “gift products with healthy margin and deep stock”, which looks nothing like any default sort.

What are we guaranteeing? Two guardrails beat ten rules: never an out of stock product in the top positions, and never a product down to its last units pushed to the top, because selling it through too quickly breaks the size run or the variant range for everything else.

What are we letting ourselves measure? Clicks by position are already in your data: if position 1 pulls ten times what position 20 pulls, moving a product up is a revenue decision. That is what turns sort order into something you manage rather than something you have opinions about.

The method: half a day per category that matters

1. Identify your top five categories by incoming sessions. That is where the work belongs: reordering a category nobody visits earns nothing.
2. Record the current order and where it comes from. Open each category in a private window and note the first ten products, their stock and their margin. It is the fastest audit in e-commerce, and the most convincing one in a management meeting.
3. Apply the rule you agreed: promoted products first, guardrails everywhere, and everything else by recent performance over a short window, thirty days rather than twelve months.
4. Check that it makes people go down. Before and after, record the scroll depth and the share of revenue coming from the first six positions. It is the only referee between an arrangement that sells and one that decorates.
5. Date it and come back to it. A sort order is a decision with a shelf life: October’s is not January’s. A twenty-minute monthly slot is enough, and it is the same habit as keeping the promotional retail calendar up to date.

Here is a five-minute self-diagnosis: open your biggest category in a private window and look at the first three products. Are they fully in stock, on healthy margin, in season? If even one of the three fails the test, the sort order is working against you right now.

Then ask the same question of a page nobody ever arranges: the results page of your on-site search, which is where your most decided visitors land.

The approaches, and what each is worth

Tooling

Reordering by hand in a shop’s admin gets tedious fast: native interfaces handle drag and drop badly, and nothing flags the out of stock items sitting at the top. That is the one real argument for a tool, and it is enough. What to look for comes down to three things: drag and drop per category, a visual alert on out of stock and remnant stock, and the ability to pin the top positions manually without the automated rule taking them back. Everything else is comfort.

The method matters more than the tool: drag and drop carries out a policy in minutes, but it is the policy, what you push and what you guarantee, that produces the revenue. A sorting tool serving no rule at all just reorders chance more comfortably.

What this says about online retail

Merchandising is the oldest job in retail, and the most forgotten one online. Entire budgets go into acquisition to bring visitors onto pages whose window display nobody has arranged: that is paying people to walk into a shop that is in a mess.

What the supermarket measurements show is more unsettling still: where architecture explains nothing because every location is equally good, all that is left is placement. A category page is that shop, taken to its extreme. Bringing that job back into the loop, driven by clicks-by-position data rather than by instinct, is one of the most profitable trade-offs we make within our e-commerce and marketplaces expertise. It also sets up the neighbouring question: a well-ordered category is the one search engines understand best, which brings you straight to the trade-off between category pages and product pages.

Sources

The verifiable claims in this article link to their primary source, consulted on 10 August 2026 and extended on 12 September 2026. We do not cite a source we have not read.

1. PrestaShop 8 documentation, Product Settings. View
2. PrestaShop 8 documentation, Browsing the front office. View
3. Google, How Search works, Ranking results. View
4. Gil, J., Tobari, E., Lemlij, M., Rose, A. and Penn, A., “The Differentiating Behaviour of Shoppers: Clustering of individual movement traces in a supermarket”, Proceedings of the 7th International Space Syntax Symposium, KTH Stockholm, 2009. Full text freely available. View
5. Hui, S. K., Inman, J. J., Huang, Y. and Suher, J., “The Effect of In-Store Travel Distance on Unplanned Spending”, Journal of Marketing, vol. 77, no. 2, 2013, pp. 1-16. View
6. Kaltcheva, V. D. and Weitz, B. A., “When Should a Retailer Create an Exciting Store Environment?”, Journal of Marketing, vol. 70, no. 1, 2006, pp. 107-118. View
7. Lunardo, R. and Mbengue, A., “Perceived control and shopping behavior”, Journal of Retailing and Consumer Services, vol. 16, no. 6, 2009, pp. 434-441. View
8. Edwards, S. M., Li, H. and Lee, J.-H., “Forced Exposure and Psychological Reactance”, Journal of Advertising, vol. 31, no. 3, 2002, pp. 83-95. View
9. Hasan, B., “Perceived irritation in online shopping: The impact of website design characteristics”, Computers in Human Behavior, vol. 54, 2015, pp. 224-230. View
10. Fortune, “IKEA tried to do away with its maze-like structure”, 1 September 2023, based on the Wall Street Journal. View

FAQ

Is arranging my category by colour harmony a mistake?

No, it is a trade-off, but you need to know which one. In a shop, colour is used to break the line and create the hook that pulls the eye further along; arranging by tonality does the opposite, smoothing the page and removing that hook. If your category page is primarily a brand surface, the trade-off holds. If it has to produce revenue, measure it: scroll depth and the share of revenue from the first six positions, before and after. If the arrangement does not make people go down, it decorates.

Does changing the product order affect SEO?

The category page keeps its URL and its content, so reordering breaks nothing. The indirect effect tends to be positive: in-stock products that get clicked at the top improve the page’s engagement signals. The one thing to watch is pagination, if products move between pages frequently.

How often should we reorder?

A monthly slot is enough in normal trading, tightened up before the big periods: October’s order prepares the fourth quarter, December’s prepares the sales. What matters is that it is a fixed appointment, not something you remember to do.

Is sorting by best sellers not good enough?

First check that it exists: on PrestaShop 8 the default sort setting offers neither sales nor popularity, and what the shop front calls “relevance” is the position inside the category. Where the sort does exist, it freezes the past: whatever sold stays at the top, so it carries on selling, including out of season or on dying stock. It also ignores margin. Used with a short window and stock guardrails it becomes usable; raw, it is just a loop.

What if my categories hold hundreds of products?

Only the first two pages deserve to be arranged by hand, because that is where nearly all the clicks are. Beyond that, an automated rule with guardrails does the job. Arranging position 150 has never sold anything.

Can Dotsland help us with this?

Yes. Reading clicks by position in your own data, writing the merchandising rule, putting the tooling and the monthly slot in place. It is one of the projects covered by our e-commerce and marketplaces expertise. Let’s talk, or start with the test on your first three products.