Collection merchandising rules that scale past a hundred products

Manual sorting is genuinely the best merchandising tool you have, right up until the catalogue outgrows the hands maintaining it. Shopify lets you drag products into order, and also move a selection to the beginning, the end, or a specific position number — which is more capable than most people realise and still does not survive contact with a large catalogue.

The arithmetic is simple. The admin lists products in pages of 60. A 400-product collection is seven screens. After a seasonal drop adds forty items and archives thirty, a correct manual order requires someone to reconsider the whole thing, and nobody does. What happens instead is that the new items get dropped at the top, last season stays at position nine, and the order becomes an archaeological record of every past intervention.

The reason this goes unnoticed for so long is that nothing breaks. The grid renders. Products appear. There is no error state for "this collection was last merchandised in March", which is why the honest question is not "is our sorting good" but "when did a person last make a decision about this collection, and would they make the same one today".

The first thing to move off human hands is which products are in the collection at all. Automated collections select by conditions, and Shopify documents the ceiling: up to 60 selection conditions per collection, with a choice of whether products must meet all of them or any of them.

Sixty is far more headroom than most merchandising needs, and reaching for all sixty is usually a symptom rather than a feature. A rule that needs thirty clauses is describing a taxonomy that does not exist yet — the honest fix is a product attribute that says directly what the thirty clauses are circling, not a longer rule.

Which is the real point: automated collections are only as good as the fields they read. If "sale" means a tag someone remembers to apply, your sale collection is exactly as reliable as that person’s memory. If it means a price comparison, it maintains itself forever. Every hour spent making product data structured and consistent pays out across every collection, every filter and every integration that reads it, and every hour spent hand-sorting pays out once.

The compromise that actually holds is to hand-place a handful of products and let the rule handle the rest. In practice the first four to eight positions decide most of what happens on a collection page; below that, ordering matters far less than filtering.

This is also the version your merchandising team can maintain. "Choose eight products a month for each of our top twelve collections" is a real job someone can do in an hour. "Keep four hundred products in the right order" is not a job, it is a wish.

Two habits make the pinning worth having. Write down why each pin exists — new range, margin, overstock, campaign — because a pin with no stated reason never gets removed. And give every pin an expiry, even an informal one. The commonest failure in a hand-merchandised collection is not a bad decision; it is a good decision from eleven months ago that nobody has revisited.

Past a hundred products, filtering does more for conversion than sorting does, because the visitor is trying to get to a subset and every extra scroll is a chance to leave. The limits are documented and worth designing around rather than discovering.

Shopify’s Search & Discovery app allows a combination of standard and custom filters "up to a maximum of 25 filters for your store", and notes that each filter source can only be used once across those filters. Standard sources include availability, category, price, product type, tags and vendor; custom filters can be built from product options, metafields and metaobjects. A single filter can display a maximum of 100 filter values, a filter group can hold up to 200 unique values, and a store can have up to 1,000 filter groups.

Two more limits matter more than any of those at scale: collections containing more than 5,000 products do not display filters, and a search returning more than 100,000 results does not either. That first number turns a vague instinct into a rule — if a collection is heading past five thousand products, it is not a collection any more, it is your catalogue with a URL, and the fix is to split it into things a person would actually ask for.

The store-wide cap of 25 is the one that changes how you plan. Filters are a shared budget across every collection, not a per-page decision, so "add a filter for this campaign" spends from the same allowance as every category on the site. Decide the set deliberately: the attributes customers narrow by most, expressed as structured product data rather than tags, and reviewed as a set rather than accumulated one request at a time.

Worth knowing before someone spends a week on it. Product boosts in Search & Discovery let you attach search terms to a product — a maximum of ten per boost — so it ranks higher for those queries, and only if the product is available for sale. Shopify documents them as affecting searches performed on your online store, with no effect on how products rank for third-party search engines.

They do not apply to collection pages. If a product needs to be higher on a category page, your tools are membership, default sort and a pinned position — not a boost. It is a small thing to get wrong and a genuinely confusing hour to spend working out why nothing moved.

All of the above assumes the collection template is a grid with a heading, which is where most of them stop and most of the opportunity goes. A large collection needs somewhere to say what the range is, how to choose within it, and which subset most people want — and that content is what makes a collection page rank as well as convert.

That means the template needs blocks a merchandiser can place: an intro with real copy, a row of sub-collection links, an editorial band partway down the grid, a guide or size chart, an FAQ at the foot. Building those as reusable sections rather than one-off templates is the difference between a collection page someone improves on a Tuesday and a ticket that waits for the next sprint.

It is also the cheapest of everything described here. Membership rules and structured filter data are real projects; giving the team a handful of composable blocks for collection pages is a small piece of theme work that pays out every week afterwards.

Most of the merchandising problems we are asked to fix on large catalogues turn out to be data problems wearing a merchandising costume. The collection is wrong because the tags are inconsistent; the filters are useless because the attribute lives in the description; the new range is invisible because the default sort rewards a decade of past orders.

Fixing the data is duller than redesigning the collection page and worth considerably more. If you only do one thing from this article, make it the default sorts — it takes an afternoon across the whole store, it changes what almost every visitor sees, and it costs nothing to reverse.

• Shopify section library
• Collection pages are your best SEO asset
• Faceted navigation without wrecking your crawl budget