As a developer-led product, Algolia is constantly expanding and improving our search capabilities. Our latest tool, Collections, enables merchandisers to curate assortments on the fly without requiring developers.

Why you need the Collections tool 

Different industries have unique requirements for grouping highly specific sets of items as part of their search strategy. Three hypothetical scenarios from three different sectors illustrate the problem:

• Fashion needs more editorial precision than broad filters allow

Premium fashion retailers like Harry Rosen need to showcase editorially curated outfits on their online stores. Before Collections, merchandisers could use the Rules context feature to filter by broad categories, such as formalwear and eveningwear, or attributes like price range and brand. But they couldn’t handpick individual items to build that perfect head to toe “date night” look: this suit, that tie, and those shoes.

• Education succeeds through structured learning paths, not broad filters

Coursera’s learning platform had a similar problem. Optimal learning progression happens by design. Coursera’s educational experts wanted to showcase sequences of 8-12 courses to optimize learning progression from beginner to expert. Rules context let course designers filter by subject area or level, but they couldn’t build themed learning paths. 

• Streamers need crafted viewing experiences to keep viewers engaged

For video streaming platforms, a binging viewer is a satisfied user. Binge-worthy experiences are a product of mood-based content creation across genres. With Rules context, content curators could filter by genre, release data, rating, and other attributes. But they couldn’t combine that one perfect comedy special with a drama series that shares the same editorial theme. 

Algolia Collections: Editorial precision without the heavy lift

Algolia Collections to solve these industry-specific requirements to showcase content beyond fixed categories. It gives content managers the ability to create dynamic, hand-picked assortments around specific themes and update them according to set conditions and criteria.

Unlike rigid categories that come from your backend catalog, Collections are defined on the fly in Algolia, letting you blend manual curation with dynamic conditions.

With Collections, Harry Rosen’s customers have access to the brand merchandiser’s expertise and vision, boosting CVR. Coursera users can take courses following a specific progression to help them land a job. Streaming platforms can carefully craft tailored viewing experiences. 

Manual curation meets dynamic conditions directly in the dashboard

Product expertise is your competitive advantage. When your product expert has a creative vision, they need a way to realize it fully, quickly, and efficiently. 

The best part of the Collections tool is that it’s built directly into the visual merchandiser dashboard. No separate tools or developer workarounds are necessary. Seamless integration with Algolia’s usual search ranking, relevance, and merchandising capabilities adds even more tailoring to specially curated content.

• Dynamic re-ranking: Automatically optimizes the order of products in a collection based on user behavior, just like with category pages today.

•  Analytics: Behind-the-scenes activity tracking lets users see how collections are performing and tweak product assortments for better results.

Everything works together to ensure tailor results reach your customers for maximum business impact. 

How do Collections work?

The previous grouping tool, ruleContexts, worked for broad personalization. Your website would send tags such as “new customer” or “mobile” along with search requests and the rules engine would filter results accordingly. It let you customize broadly, but merchandising and ecommerce experts couldn’t handpick individual items or create nuanced collections. 

With the Collections update, when a business user creates or modifies a collection in their dashboard, this triggers Algolia’s transformation engine. The engine evaluates your collection conditions both for dynamic rules and any handpicked items.

At this stage, the transformation engine also adds the ‘_collections’ attribute to any matching records, which adds the transformed records to your index. Your search index now contains those products that are properly tagged with collection membership.

In other words, whenever a new product enters the catalog or index that matches your collection conditions, it’s instantly tagged. If it stops matching, the tag gets removed so collections stay in sync without manual effort.

With Rules contexts, rules are applied when something is searched for. With Collections, we've pre-computed collection membership and stored that as data. This is a far cleaner and more reliable method that enables precise expert curation. Everything flows through the same transformation pipeline, keeping collections-related data accurate and up to date.

There’s simplicity on the frontend, too. Under the hood, Collections work exactly like any other facet in Algolia. When a customer visits a collections page, search results are filtered by the ‘_collections’ attribute that was added during indexing, just like any other facet in InstantSearch.

Here’s what the code looks like in Vanilla JavaScript in InstantSearch. To set up Collections, create a menu widget that filters on the ‘_collections’ attribute.

There’s no need to create new pages for every single collection. One dynamic template handles all your collections and works without routing. As a result, you get clean URLs, such as collection/datenight for Harry Rosen’s editorial looks.

This means minimal work for developers. Once the initial setup is complete, business users can build and swap out items in their collections without waiting for any additional coding from their dev team.

What’s next for Collections 

With AI-powered capabilities, there’s potential to do even more with Collections. We’ve already launched a few new features. Here’s what’s available now:

• Enhanced indexing support with flexible conditions: Expanded filtering capabilities will support numeric operators and complex logic, delivering more precision and control. You’ll be able to filter for “price under $100” and combine filters using “and/or”.

• Smart Group integration: Smart Groups let you define curated sets of products within search results or category pages, such as a “sponsored product” block. You can now inject Smart Groups into your collections pages, too. Smart Groups stay consistent across search category pages and collections pages for a unified merchandising strategy.

• Neural Collections: Generate new collections in seconds by typing a natural language prompt. Review matching products, tweak, and scale as needed, saving time and effort.

Higher Collections limits coming soon. Algolia Premium and Elevate users can currently create 25 1000-item collections per index. Starting on September 2nd,  You’ll be able to create 1000 collections containing over 2000 items.

Leverage your business users’ expertise to design campaigns and seasonal promotions without waiting for developer capacity. 

What makes Collections special is its blend of human manual control and dynamic AI curation. Your merchandiser intent is transformed into a working user experience instantly, while maintaining technical integrity and performance. 

To learn more about Algolia Collections, watch my full DevBit video:

Or get in touch with our team to get a demo.