We build catalog search, so we have an interest in what follows. We have tried to make it useful anyway — the method is stated, the limitations are stated, and where a competitor is the right answer we say so.
We ran the same queries at nearly thirty distributors. What we found: zero-results pages on one wrong letter at a handful of distributors, national names among them, result counts collapsing from 121 to 2, spec queries ignored, accessories outranking the product — and the finding we did not expect: paying for a search vendor did not predict having good search.
Read it →Algolia, Coveo, HawkSearch, Klevu, Searchspring, Constructor, Elasticsearch, Solr, Typesense, Meilisearch and ours. Written by a company that sells one of them, which we say up front — and several sections end by telling you to buy someone else.
Read it →The subscription is the number everyone compares, and it is rarely the biggest one. A five-year model covering infrastructure, integration, the relevance maintenance nobody books, and the largest cost of all — the customer who could not find the product and left.
Read it →