Adaptive search for easy and fast discovery

Search on a large webshop site usually returns a lot of unwanted results. Especially searching with general terms when you don’t know or remember exactly what you are looking for (i.a. ”shirt”, ”shoes”). In this example the user is looking for a pair of sneakers. Searching with ”sneakers” in Amazon returns over 50 000 results. And specifying your search isn’t very easy.

An adaptive search would make the next steps toward discovering the correct product easier and faster.  Each item on the search results would have 2 options: remove an item from results and narrow results down to certain items.

The webshop platfrom knows a lot about the product: category (shoes), subcategory (sneakers), target group (men), brand (xyz), color (red), price, rating and probably a lot more. The 2 options, ”remove” and ”narrow down”, would use this information to create a new search with new parameters. For example clicking a red sneaker would show: ”remove all non-red sneakers” or ”show only red sneakers”.

When the user clicks the ”remove” option on an unwanted search result, the system would show the characteristics of that item and the user could easily select which ones are incorrect or unwanted. And if the product is almost the right one, selecting it would be a step to the right way of finding the desired product. So clicking the ”narrow down” option would  show the characteristics of that product – as with the ”remove” option. The user could easily select which characteristics are correct.


– the user could start searching fast without inventing up a long, precise search phrase

– the user doesn’t have to learn how the search engine works and find out if he should use ”-black” or ”NOT black” or ”+red” or what. The platform creates new search phrases

– less frustration

– more sales


I’m using Amazon as an example site, because besides being a mess, it is  the #1 webshop and has the biggest selection of products and would really benefit from a better search.

Let’s say you’re looking for shoes you saw in a store / ad few days ago. You have no idea of the brand, but you remember they were men’s fashion sneakers, ”reddish” and slip ons. Price was about $100.

You type ”sneakers” to the search, hit enter and you get a looooong list of results. You don’t see what you’re looking for – some are in the right product group and some clearly are not. Now what? You start to look for something to redefine the search – eventhough you can clearly see what are closer to what you are looking for and what are not. If there was a sales clerk in front of you, you could say ”they were not for running and almost like those Pumas. .” And then the clerk would get new sneakers more like the Pumas.

An adaptive search could do the same – the two options (”remove” & ”narrow down”) are shown. You select ”narrow down” (or ”+” in the image) and select characteristics that match what you are looking for:

Now you have a new list of sneakers that all look more like the one you are looking for. But still, there are a lot of wrong ones. So, by clicking one of them gives a list of its characteristics and you can choose which characteristics don’t match the product you are looking for:

Now you are very close! The search filters the user have created with his selections should be visible above the search results and other, still unused, but useful filters should also be placed there. Like colors in the image below. The user don’t have to click through all products that are the wrong color – he just clicks red. (Note: color selection is possible at Amazon).

And there are the shoes!! Just a few clicks and no wondering how the search works or where are the correct buttons. Easy and fast!