The 7 Types of Ecommerce Search (and Why Most Sites Still Get Them Wrong) 

Ecommerce customers search in seven main ways: exact, product-type, feature, use-case, compatibility, problem-based and non-product (information) searches. Most sites handle exact searches well but fail on the others, often because product data is incomplete. Fixing your catalogue data first, then adding semantic or AI search, closes most gaps. 

Baymard's ecommerce search benchmark of over 170 shop sites and apps determined that 56% did not meet users' search needs well. It also discovered that about half of users use search as their first port of call for a product.

The 7 Ways Customers Search Your Store 

A customer is looking for a product on your ecommerce website. The product is available at a suitable price and is listed in your catalogue. But in the end, they didn't find anything useful. 

This may not be a traffic issue. It may be the site's search experience. 

Site search matters because it serves shoppers who already know roughly what they want, and those are your highest-intent visitors. 

Here are the seven types of searches: 

1. Exact Search: “I Know What I Want” 

Some customers know exactly what they want. They can use product name, model number, SKU or brand to search. 

A customer could type “Sony WH-1000XM6”, “Nike Air Max 270” or “Dell XPS 13 charger”. These searches should be relatively simple as the clue for identifying the product has been provided by the customer. 

However, Baymard also discovered that 12% of websites have problems supporting exact searches. 

The reason may be due to the discrepancy between the customer language and data in the catalogue. Customers often type the manufacturer’s product name, while your catalogue uses a shorter one., while your catalogue refers to it by a different, shortened name. 

Good ecommerce search should handle:

• Spelling errors
• Product naming
• Alternative names to a product
• SKUs and model numbers
• Common abbreviations 

When a customer queries for a product you have in stock and gets a “No results”, they may just think you don't have it in stock.

Every failed search is a lost sale, which is why search is one of the first things we check in conversion rate optimisation work

2. Product-Type Search: “Show Me This Kind of Product” 

Not all customers have a definite target in mind. At times they will only know the category.

These could be “women's trainers”, “office chairs”, “wireless headphones” or “coffee machines”. The idea is that the search engine can glean the type of the product being searched and present the user with a relevant set of products. 

Baymard discovered that 20% of sites were also experiencing product-type search problems. 

This is a bigger problem on large catalogues, where thousands of products share the same keywords.

Search should be able to comprehend product, product category and product sub category. It guides customers through the process of a general-based search to the appropriate product category without forcing them to be familiar with the retailer's categorization structure. 

3. Feature Search: “I Want This, With These Requirements” 

Customers frequently select a specific product type and have specific requirements. Their choice of words may be “blue linen shirt”, “256GB phone”, “£100 waterproof jacket” or “size 8 running shoes”. 

These searches are more complex, as there is a product description and the features that are desired. For attributes of a product, it is important to describe them rather consistently throughout the description, such as colour, size, material, capacity, price, compatibility, technical specifications, and so forth. 

For an online retailer, for example, when a user searches for ‘blue shirts', the machine should know that "blue" is a colour requirement. It should then suggest products suitable to the customer's needs, and let them narrow the results down further. 

While AI can help in making this process more natural, good search begins with good catalogue information. 

4. Use-Case Search: “I Need Something For This Situation” 

Customers don't always look for products by name. They tend to talk about what they would like to achieve. 

The customer could be looking for “Laptop for graphic design,” “dress for a summer wedding,” “Furniture for a small flat,” or “Shoes for long-distance running.” 

These aren't just product categories, they are a use case. 

Baymard found that 43% of sites had issues supporting use-case searches. 

You need to know how customers actually use your products. A jacket may be labeled waterproof, insulated, and lightweight, but if these terms aren't correlated in the catalogue with hiking, they probably won't be able to find it by performing a search on its name. 

Use-case information can be added to product attributes, category content and merchandising rules. However, a customer can use a natural language query and then a semantic search pulls the relevant product from the results. 

5. Compatible search: “Will this fit into the rest of my tools?” 

Compatibility search is absolutely essential for electronics, automotive parts, appliances and technical products. 

The customer can look up “Dyson V11 filter”, "a Canon EOS R6 battery" or a "BMW 3 Series brake pad". They are familiar with the product they currently have in possession but possibly not the actual name of the part they desire. 

Baymard found that 44% of sites had issues supporting compatibility searches. 

This can have a direct impact on returns and customer trust. A recommendation of an incompatible accessory or component leads to returns and lost trust.

If you sell technical products, compatibility data is essential.

6. Problem-Based Search: “I Have This Problem” 

A customer may not have exactly what he or she needs to solve a problem. Rather they look for the problem. 

Such as 'slow laptop ', 'stain on white sofa', 'leaking kitchen tap', or 'garden weeds'. 

This is a chance to bridge the gap between search and products, and the advice and recommendations received to the bought items. A home improvement store might link the work of “leaking tap” to the right kind of repair products and offer a troubleshooting document. 

The ability to do AI search can be of great help here, as it can match a described problem to the right products, even when the customer doesn’t know the product name.

7. Non-Product Search: “I Need Information” 

That is not the only purpose you can put your ecommerce search bar to. 

Other queries customers look for include: 

• Return policy
• Delivery information
• Refund
• Warranty
• Order tracking 

According to Baymard, 34% of users tried to find something other than a product, and the new benchmark statistic of Baymard has revealed that 15% of the ecommerce websites they reviewed did not allow basic searches that users would need for anything but a search for a product. 

If someone searches ‘returns’ and gets product listings, your search is broken.

Not sure which of these seven searches your store gets wrong? Ecommerce Excellence can audit your site's search and product data and show you where customers are leaving. 

Book a site search review. 

The Practical AI Opportunity 

The best use of AI-powered ecommerce search is to handle complex, multi-requirement queries like this:

“I need a laptop for my small design team, costing around £1,500, that can run Adobe Creative Cloud applications.“ 

Traditional keyword search struggles with this mix of budget, context and performance needs. AI search can also understand the query, highlight key product specifications and deliver appropriate product results, along with describing how it accomplishes this. 

The secret lies in maintaining the system as grounded in actual catalogue information. Don't let AI create products, product specs or availability! 

AI shouldn’t make search more complex. It should make it better at understanding customers.