1. What a site search has to handle
Customers type short and often imprecise queries: a product name with a typo, an article number, a brand, a colour plus a product type, or a word your catalogue does not use. A good site search finds the right products for all of these, ranks available products sensibly and offers filters that match what customers actually narrow down by.
The built-in search of many shop systems only looks at a few text fields. SKUs, attributes and synonyms are often missing, which is why stores add a dedicated search. Our guides on the default search of WooCommerce, Shopify, Magento and PrestaShop show what each one covers.
2. Build a test set from real queries
Before you compare providers, write down around 20 queries your customers really use. Your shop statistics, Google Search Console or customer service emails are good sources. Run the same list in every search you test and note whether the right product appears in the first few results.
- Three article numbers or GTINs, once complete and once with a part missing.
- Three product names with a typical typo.
- Three queries with an attribute, for example “red dress” or “jacket size L”.
- Three words your catalogue does not use, for example “couch” if your products say “sofa”.
- Three brand names, singular and plural forms, and one query for a sold out product.
- A few queries that currently return no results in your store.
3. Product data decides more than the algorithm
A search can only find and filter what is in your product data. If colour, size or material are missing in the feed or the store, no provider can offer a reliable filter for them. If the SKU is not exported, customers cannot search for it.
So check your data first: are SKU or GTIN, brand, category and the attributes customers filter by filled for most products? Many problems that look like a weak search are gaps in the data. Our product feed guide lists the fields that matter.
4. What you should be able to control
Results will never be perfect out of the box, so look at what you can adjust yourself: synonyms, which fields are searchable, which attributes become filters, pinning a product to the top for a query and ranking rules, for example placing available products first.
Also check the reports. A list of frequent searches without results is the most useful starting point for synonyms and data fixes, because it shows what customers look for and do not find.
5. Compare pricing, installation and data location
Providers count different things. Some bill by requests, where a single search can create several requests while the customer types; others bill by searches or by the number of products. Estimate your monthly searches and catalogue size and calculate the price for each model, including what happens when you exceed a limit.
Installation differs as well: a native app or plugin, or a product feed plus a small script in the store theme. Ask how often the product data is updated, whether prices and stock are current and how the search looks on mobile. For stores in the UK and the EU, check where the data is processed and whether the search sets cookies.
6. Where Grumeria fits
Grumeria is a site search for small and mid-sized stores. It includes typo tolerance, synonyms, filters, ranking rules and statistics in every plan, with suggestions from frequent searches without results. It does not offer visual search or 1:1 personalisation. Plans start at €29 per month excl. VAT for 2,000 products, with a 14-day trial without a payment method. Search runs in Frankfurt without cookies.
Shopify, WooCommerce, Magento 2 and PrestaShop stores can usually start with just the store address; other systems such as ShopWired or BigCommerce use a product feed and a script snippet. The free feed check shows in one minute which data Grumeria finds and runs three example searches on your products.