How Does Etsy Search Work?
When a user searches for an item on Etsy, Etsy’s search engine works to return the most relevant matches. This is accomplished in two phases for organic search results. First, Etsy search scans keywords across titles, attributes (e.g. color), categories, and tags in each product listing related to the search terms (a query) and returns an initial set of results. This phase is known as query matching.
Second, once Etsy search gathers all listings related to a buyer’s query, we use the information we have about each listing and shop to rank the listings so people browsing Etsy are shown items that they are most likely to purchase earlier in the results. This is the second phase, which we refer to as ranking. There are a number of different factors that are used when ranking listings in search results. Each is weighted differently and impacts ranking in organic search results. The most important factors are discussed below:
- Relevancy: relevance refers to how proximate keywords in a listings’ tags, titles, descriptions, or attributes are relative to the search query a member is searching for.
- Listing Quality Score: to determine a listing’s quality score, we factor in a combination of data such as clicks, views, favorites, and purchases to determine how well the listing will convert into a purchase or may be relevant to shoppers. The higher the quality score the higher the item may appear in search results.
- Recency: how recently an item was listed or re-listed by a seller. When a new listing is created, it gets a small, temporary boost in search results so that Etsy search can learn more about how buyers interact with it to determine the listing’s quality score. Renewed listings get a similar, though smaller, boost.
- Customer & Market Experience Score: to determine a shop’s Customer & Market Experience Score, Etsy considers a variety of factors that make up overall shop quality, including user reviews of products, whether or not a shop has completed an about section or its shop policies (e.g. returns), whether or not the shop may be able to fulfill orders in a timely fashion, and whether or not the shop may have violated any Etsy policies.
- Shipping Price: shipping price is a factor in search ranking. If sellers offer a US free shipping guarantee or free shipping to US shoppers or if they offer lower shipping costs to buyers in their own country, it may improve their listings rank in search results.
- Translations, Location, and Language: the titles and tags sellers add to their listings are typically in the language chosen during shop setup or based on their preference. Etsy translates seller keywords and serves buyers results that match their query and language settings. In certain markets, Etsy search also prioritizes ranking local shops in search results.
- Frequency Capping: to limit the number of times the same listing or shop appears at the top of results for a specific user or keyword.
- Context Specific Ranking: when available, search results are customized to individual buyers based on Etsy’s Context Specific Ranking technology, which takes into account, among other factors, related keywords and what buyers are interested in and likely to purchase.
To learn more about how Etsy search works, including why Etsy considers the above ranking factors relevant, check out this Help Center article and our Ultimate Guide to Search.
Etsy may highlight a cross-section of shops or listings organically, above or alongside primary search results, based on some or all of the factors above.
Etsy does not offer sellers a higher ranking placement in organic search results in exchange for any compensation or fee. Sellers can, however, purchase on-site ads, called Etsy Ads, which are identified separately as ads on Etsy. This article is incorporated into and is a part of our Terms of Use and Seller Policy.
How do Etsy Ads work?
Etsy Ads are based on an auction for ad space on Etsy’s properties. Etsy enters "bids" for this ad space on a seller’s behalf based on a daily budget set by the seller. Specifically, Etsy Ads use a sealed bid generalized second-price auction system to determine the cost and placement of ads. In other words, ad inventory is "auctioned off" at varying prices, depending on the competition related to a query. For example, if more Etsy Ads qualify for a particular search query, the auction price may also increase. A bid is the maximum amount of money a seller will be charged if a buyer clicks on the seller’s Etsy Ad. Etsy’s calculation of bids is based on information provided, such as the seller’s daily budget or the listings the seller wishes to include for Etsy Ads. It may also be based on other information collected by Etsy, such as, for example, search relevancy, listing variety, listing quality (including seller compliance with Etsy's policies), or likelihood of views or sale. For more information on the factors that go into Etsy search relevancy, see Etsy’s Ultimate Guide to Search.
For each Etsy Ad, Etsy's bid management system will automatically determine a seller’s bid selection ("auto-bidding"). Auto-bidding is designed to determine the bid price that will deliver positive return on ad spend for a particular listing within a given budget over a day, where those auto-bids are subject to the auction process described above.
Etsy Ads are distinguished to users from organic content by being marked or referred to as seller ads, sponsored content, promoted listings, or using other distinguishing language, at the listing level or for a designated content area. To learn more about how on site ads work and how ads are ranked, check out our Help Center and Etsy’s Advertising and Marketing Policy. This article is incorporated into and a part of our Terms of Use and Advertising and Marketing Policy.
How do Etsy Recommendations work?
Recommendations help users find a special item among more than 100 million unique listings on Etsy. Recommendations appear at multiple touchpoints across Etsy, and depend on a user's shopping mission – whether browsing trending items, picking up shopping where they left off, or discovering new content and interests based on their prior activity.
We call each recommendation set a “module,” and there are hundreds of modules both on the web and on mobile apps. Recommendation modules work in two phases: selection, which retrieves a small set of relevant items out of the entire inventory as quickly as possible, and ranking, which ranks the items using a more sophisticated machine learning model, typically with an emphasis on the user’s current shopping mission, and decides on the best few items to offer as recommendations.
There are a number of different factors that are used when ranking recommendations. These models are called “rankers.” Rankers score items for relevance based on both contextual features and item features. Rankers are optimized based on feedback from a user’s engagement with the recommendation modules. While the criteria vary, the most significant criteria used to determine the recommendations are:
- Listing features including data such as price, location, shipping and shipping price, historical interaction, images, attributes, shop quality
- User features including data such as location, shipping, historical and real time interactions, search queries, hobbies and interests, textual features
- Real time features including data such as recently viewed items, recent search queries, recently searched taxonomies
- Time-based features including data such as day of the week, hour of the day, month of the year
The relative importance of those features in the recommendation module is weighed based on the next best action for a buyer - whether browsing, continuing a shopping mission, or ready to purchase.
There are certain recommendation modules that are dedicated to curated content, like “Shop the Look,” which identifies objects in an image and finds visually similar listings; “Gift Mode” which uses machine learning to identify listings based on their relevance to a broad category persona (e.g., “The Musician”) or gift idea, along with their historical performance on Etsy, their likeliness to be a gift, and their Listing Quality (as defined above); and other collections such as holiday gifting events, sales, and other collections created by Etsy editors.
Etsy also uses its recommendation systems to enable in-app updates and notifications, which inform users when items have gone on sale, a seller is offering new items, or other updates related to a user’s activity. Finally, Etsy recommendation systems use query ranking, using past listing engagement to determine relevant queries to surface.
The most effective way for a user to modify the recommendations and order is by clicking listings, favoriting (or un-favoriting) an item or a shop, which is an option available in the recommendation module. A user may also clear recently viewed listings in the privacy tab of their profile and update their privacy settings. Most recommendation modules have a recency bias, so a user may also influence the recommendation modules through searches, clicks, purchases and other activity.
Last updated on Feb 16, 2024