"I can't find it" and "we don't have it" end the same way: the shopper leaves without buying. The difference is that in the first case, the product was there all along.

It happens more than you'd think. The Baymard Institute, which has spent years watching real people shop online, has seen participants abandon stores again and again because they couldn't find what they were after, sometimes with the product right there on the site and a search box that never led them to it.

This is what's known as product discovery: everything between someone arriving in your store and seeing the product that suits them. And it isn't one problem, because some shoppers arrive knowing far more than others.

Three shoppers, three different problems

The one who knows exactly what they want. They search for "Asics Gel-Nimbus 27, size 9". This is the easy case: in Baymard's 2026 analysis, only 12% of stores had trouble with exact searches.

The one who knows the type of product, but not which one. They search for "road running shoes" or "light waterproof jacket". This is where it gets harder. 39% of stores struggle with searches by feature ("waterproof", "light") and 43% with searches by use ("for roads", "for travelling").

The one who doesn't know what they want. They have a problem or an occasion: "something for back pain when working from home", "a present for my dad, who cycles". Baymard found 37% of stores struggle with searches that describe a problem.

The less a shopper knows when they arrive, the harder discovery becomes. And that shopper, the one who hasn't decided yet, is exactly the one who most needs the store's help.

Where shoppers get lost

Almost always in one of four places:

  • Categories. They're organised the way the store thinks, not the way the shopper does. You know the hydration packs live under "Trail accessories"; they look under "Backpacks". In Baymard's benchmark of 343 large US and European stores, 76% were mediocre or worse at category navigation.
  • Search. Many search boxes only find what matches your product pages word for word. If the shopper types "sweatshirt" and you call it a "hoodie", they get nothing back.
  • Filters. They filter by what's easy to measure (colour, size, brand), not by what the shopper uses to decide (what it's for, in what conditions).
  • Recommendations. The usual "you might also like", showing products similar to the one the shopper is already looking at, when they needed something different.

What physical shops still do better

In a Salsify survey of more than 2,700 shoppers in the US, Canada and the UK, physical shops came out as the top place people discover products (60%), ahead of online marketplaces (57%). Respondents said discovering products in person gives them more confidence in their quality than any digital channel.

It's not hard to see why. In a shop you can say "I'm looking for something for this" and someone takes you to the right shelf, even if you didn't know what it was called. Online, the shopper has to translate their need into the words your search box understands or the structure of your menu.

How to test your store in ten minutes

Before changing anything, try this. Pick a product that sells well and look for it in your own store in five ways:

  1. By its exact name.
  2. By the type of product, using a different word ("sweatshirt" if you say "hoodie").
  3. By a feature ("waterproof", "cordless").
  4. By what it's for ("for travelling", "for a small flat").
  5. By the problem it solves ("to keep my coffee hot").

Every search that doesn't bring the product up in the first few results is a shopper who would leave without seeing it. Repeat the test with two or three more products and you'll see a pattern.

What to improve, in order

  1. Your product pages. Search, filters and any AI tool all work from your product data. If nothing on the page says it's waterproof or what it's for, it's hard for any tool to find it that way. How to prepare your catalogue for AI has a practical checklist.
  2. Your customers' words. Look at what people search for in your store and use their words in categories, filters and search synonyms.
  3. A search that understands meaning. That's what semantic search does, and it makes a big difference to searches by use or by problem.
  4. A way to describe the need. For the shopper who doesn't know what they want, the closest thing to a physical shop is being able to explain it in their own words and get a recommendation. That's what an AI shopping assistant such as SmartShop AI does with your catalogue.

Discovering a product is only the first step; next comes choosing between the ones that fit, which is a different problem and one I cover in how to help customers choose. But if the shopper never sees the product, nothing else matters.