How to use Sorting Optimization
Default sort order quietly decides which products your shoppers see first — and which ever sell. This guide shows how to pick, test, and personalize PLP sorting for revenue.
Sorting Optimization
Choosing and tuning the default sort order on category and search pages so the most profitable products get the most visibility.
Sorting optimization is the practice of picking, testing, and personalizing the order in which products appear on a product listing page (PLP) before the shopper touches a filter. The default sort — best-sellers, newest, price low-to-high, relevance, or a custom merchandising score — controls the first ten to twenty tiles a visitor ever sees, which is where most clicks and revenue come from.
It sits inside the broader discipline of PLP optimization, alongside filters, faceted navigation, and tile design. Get it wrong and you bury your margin-rich SKUs under loss-leaders. Get it right and you lift revenue per visitor without acquiring a single new shopper.
On a typical category page, the first row of products receives 40-60% of all tile clicks. The next row gets roughly half of that. By the time a shopper reaches row five they've clicked on something, applied a filter, or bounced. Whatever sits in those top rows by default is, in effect, the store you've chosen to show.
That makes the default sort one of the highest-leverage settings in the catalog. It's also one of the least examined: most stores inherit whatever their platform shipped with — Shopify's manual order, Woo's date-descending — and never revisit it as the assortment grows from 50 SKUs to 500.
Why the default sort decides the page
Shoppers anchor to position. Eye-tracking and click-map studies consistently show an F-pattern over PLPs: the top-left tile gets the most attention, the top row dominates, and engagement decays sharply below the fold. Whatever sits in slot one is treated as the store's recommendation, whether you meant it that way or not.
Only a small minority of visitors change the sort manually — in our experience around 5-12% on desktop and even fewer on mobile, where the dropdown is hidden behind a sheet. Filters get used more, but they re-rank within the current sort, not against it. The default is what 88-95% of sessions actually see.
The compounding effect matters too. The products you surface sell more, which boosts their best-seller score, which surfaces them again the next day. A weak default doesn't just under-perform on day one; it trains the algorithm on the wrong winners for the rest of the quarter.
Best-sellers is not always the best default
A pure best-sellers sort on a store with skewed inventory will keep showing the same 8-12 SKUs and starve the long tail. New launches never get oxygen, sale items get over-exposed, and conversion plateaus because nothing in the lineup feels fresh on a return visit.
Choosing the right default sort
There's no universal winner. The right default depends on assortment size, refresh rate, margin spread, and the intent of the category. A small apparel store with 60 SKUs and seasonal drops behaves nothing like a 5,000-SKU electronics catalog with stable pricing.
As a starting heuristic: use best-sellers when your top SKUs convert above category average and you have enough breadth that the top 20 won't look repetitive. Use newness for fashion, beauty, and any category where returning shoppers expect novelty. Use a blended merchandising score — popularity weighted by margin, stock health, and recency — once you've outgrown the simple options.
Revenue per visitor by default sort (indexed, blended sort = 100)
Price low-to-high is the trap default. It anchors shoppers on the cheapest item, pulls AOV down, and starves your margin SKUs of exposure. Keep it available as a user-selectable option — many shoppers want it — but rarely make it the default outside of clearance pages.
Measuring the impact of a sort change
Treat default sort like any other PLP variable: run it as an A/B test, not a gut call. Split traffic at the category level, hold the assortment and filters constant, and watch revenue per visitor as your primary metric. PLP click-through rate is a useful guardrail but it can rise while RPV falls if the new sort pulls shoppers toward cheaper tiles.
Two weeks is usually the minimum readable window — long enough to ride out a weekday/weekend cycle and any paid-traffic spikes. Stratify by device, because mobile shoppers behave very differently from desktop on PLPs, and by category depth, since the effect of a sort change scales with how many products sit below the fold.
Typical PLP metrics by default sort strategy (online retail, AOV €40-120)
| Default sort | PLP CTR | Revenue per visitor | Avg. order value | Notes |
|---|---|---|---|---|
| Manual / platform default | 12-16% | €1.40-1.80 | €62 | Stale; degrades as catalog grows |
| Newest first | 14-18% | €1.60-2.00 | €68 | Strong for fashion, beauty |
| Best-sellers | 16-20% | €1.80-2.20 | €71 | Risk of top-SKU concentration |
| Price low-to-high | 15-19% | €1.20-1.60 | €48 | Pulls AOV down; avoid as default |
| Relevance (search PLP) | 18-22% | €1.90-2.30 | €69 | Best for search results, not browse |
| Blended merch score | 17-21% | €2.00-2.40 | €73 | Highest ceiling; needs tuning |
If you imported historical GA4 data when setting up analytics, you can backfill what each category's RPV looked like under the old sort before you ship a test — useful for sanity-checking that your A/B result isn't being skewed by a seasonal swing or a paid campaign you forgot was running.
Personalization, SEO, and guardrails
Once a global default is performing, the next lever is per-segment sorting. Return visitors don't need to see the same top row they saw last week; show them newness instead. Paid social traffic landing on a category often converts better with a best-sellers default because intent is shallower. First-time organic visitors usually want relevance.
Watch the SEO guardrail carefully. If the default sort changes per session, Googlebot still needs a deterministic crawl version of each category URL — otherwise rankings drift as content shuffles. The standard fix is to serve bots the canonical sort and personalize only the rendered DOM for known sessions.
Audit before you optimize
Pull the last 90 days of category-level data and rank pages by traffic × bounce rate × low add-to-cart rate. Those are your sort-optimization candidates. A category doing 8% of sessions and 3% of revenue is screaming at you that the default sort is hiding what shoppers actually want.
Frequently asked questions
There's no single answer, but most Shopify stores over 100 SKUs outperform the platform's manual default by switching to best-sellers or a blended merchandising score. Smaller stores with frequent drops often do better with newest-first. Run a two-week split test before committing.
Daily for fast-moving categories like beauty and fashion, weekly for stable categories like home goods or electronics. A best-sellers sort that refreshes only monthly will keep promoting SKUs that have already peaked, and it'll miss whatever started selling last week.
Not if the canonical URL keeps rendering a deterministic order for crawlers. Problems start when each session gets a different sort and Google indexes inconsistent content. Serve bots one stable sort, personalize for logged-in or returning users in the DOM only.
Usually yes, for consistency, but the impact size differs. Mobile shoppers see fewer products above the fold, so a bad default punishes them harder. If you have to optimize one device first, optimize mobile — it's where the majority of PLP traffic now lands.
Sorting decides the order of products in view; filtering and faceting decide which products are in view at all. Sorting affects everyone who lands on the PLP; filters only affect the 20-30% of shoppers who interact with them. Sorting is the bigger lever for that reason.
A custom rank that combines several signals — typically 30-day units sold, gross margin, stock level, and freshness — into a single number per SKU. It's the most flexible default sort because you can dial each weight up or down to match the category's commercial goal.
Sort starts to matter at around 12-15 products — roughly two scrolls on mobile. Below that, every product is essentially visible. Above 30 products, the default sort is the single biggest determinant of which SKUs ever get a click.
You shouldn't sort purely by margin — shoppers will feel the disconnect from popular taste. But weighting margin as one input in a blended score (say 20-30%) is standard and ethical. It nudges the top of the page toward your profitable SKUs without burying what people actually want.
Split traffic 50/50 at the category template level, hold filters and assortment constant, and measure revenue per visitor over at least two full weeks. Watch AOV and add-to-cart rate as guardrails. Stratify results by device and traffic source before you call a winner.
Audit your three highest-traffic categories. If any are still on the platform's inherited default, switch them to best-sellers and re-test in two weeks. That single change is the most common 3-7% revenue lift we see on stores that haven't touched their PLP settings since launch.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.