Algorithmic ranking The process by which a search engine orders results based on a combination of textual relevance, business signals, and user behavior. Unlike manual sorting or fixed rules, algorithmic ranking adjusts dynamically, weighing multiple factors simultaneously to surface the most relevant result for each query. Definition Definition Algorithmic ranking is the overall process behind how a search engine decides which order to show results in. Instead of a person sorting products by hand, or using one fixed rule ("always sort by price"), the system weighs several factors at once, textual relevance, business signals, and user behavior, and calculates a position for every result based on all of them together. Textual relevance covers how closely a query matches a product's title, description, and tags. Business signals cover things a store cares about independent of the query, like stock availability, margin, or a manual boost on a specific item. User behavior covers what tends to happen when real shoppers see this product for this kind of search: click rate, add-to-cart rate, past conversion. None of these factors work in isolation. The system combines them into a single score per product, then orders results by that score. This is different from manual sorting, where someone picks the order directly ("show these three products first"), and different from a fixed rule, where the same logic always applies regardless of context ("sort alphabetically" or "sort by price, low to high"). Algorithmic ranking adjusts based on the situation: the same catalog can produce a different order for a different query, a different shopper, or a different moment, because the underlying signals feeding the score keep changing. What it's used for What it's used for Without an algorithmic approach, a store either has to sort results using one rigid rule that ignores everything else relevant, or has someone manually decide the order for every possible search term, which doesn't scale past a small catalog. Algorithmic ranking lets the order reflect more than one thing that matters at the same time. A product can be a strong text match and still rank appropriately lower if it's out of stock, or rank appropriately higher if shoppers consistently choose it whenever it shows up for a given search. That produces results people actually want to click, not just results that technically contain the right words. How Doofinder applies it Doofinder's search runs on an algorithmic ranking model by default: every result gets a relevance score based on textual match, and store owners can layer in Relevance Criteria on top, additional fields like stock, price, or a custom attribute, that act as ordered tiebreakers when relevance scores are close. Store owners configure this under Search > Advanced Preferences > Relevance Criteria in the admin panel, choosing which fields matter and in what order they should apply after the base relevance score. Example Case study A furniture store sells several similar dining tables that all match the search "oak dining table" almost equally well by text. Without algorithmic ranking, all three would show up in roughly the same position, decided by minor wording differences in their descriptions, regardless of which one shoppers actually want. After the store sets stock level and past conversion rate as ranking criteria alongside text relevance, the table that's in stock and sells consistently well moves to the top, a similar table that's low on stock drops slightly, and an older listing that rarely converts settles toward the bottom. The order still respects the original text match, but now accounts for what's actually available and what shoppers actually choose. References Sources https://support.doofinder.com/search/search-setup/relevance-criteria