As-you-type autocomplete is a well known search pattern that many of the best websites offer. Providing suggestions helps the user enter longer queries. For mobile users, it reduces the amount of typing required to execute a search. But, it’s not just about typing less. It’s far more powerful — it helps users formulate the best search. Good suggestions are proven queries that ensure the best results. The quality of the suggestions is as important as the results they generate.
Today, we’re releasing Query Suggestions, a new feature that extends the conversational search experience. It is as if the user, with every keystroke, is conversing with the search bar, discussing with it exactly what to send to the search engine to obtain the best results. For this to work, suggestions need to be instantly responsive and intuitively/accurately relevant to be useful. If either one of these is missing, the conversation is over.
While generating suggestions might appear trivial, it comes with its own set of challenges. Suggestions often derive from popular queries: pick what your users have been searching for most, and suggest it next time so they don’t have to type it.
The popularity of each query, extracted from analytics, can also be used as a great way to rank suggestions. If “iphone charger” is the most searched term, it might make sense to make it the first suggestion when your users are typing “iph” — but not always.
Showing only relevant suggestions
Not everything your users are searching for should end up as a suggestion. As is often the case when it comes to data, it needs to be processed and cleaned. For example, it would be a mistake to suggest a query that leads to no results. This is where our Query Suggestions tool comes into play. You are now able to configure a set of rules to create the most relevant suggestions for your business.
Preventing spam & cheaters
One caveat to using analytics as a source of suggestions is relying on user-generated content: user input should never be trusted blindly.
Castorama, a French home improvement tools and supplies retailer, made that mistake last year. Typing any kind of a query multiple times was the only requirement for it to start appearing in the autocomplete for everyone. Their system got flooded with inappropriate suggestions and Castorama had to shut down their website for weeks, missing many of sales opportunities.
For marketplaces, it is equally important that vendors cannot influence how suggestions are ranked; otherwise, they might be tempted to promote suggestions that favor their own products.
Introducing Query Suggestions
Our Query Suggestions tool has been created to address these issues. Popular queries will be extracted from our Analytics API, and then processed and cleaned. While we have systems in place to limit inappropriate suggestions and prevent fraud, suggestions can also be moderated by the user with an easy-to-use blacklist.
We also support external analytics data, which will be particularly useful for new customers who can upload any historical data from previous search solutions, or Google analytics. And for more complex use cases, combining facets is also a possibility: for example a brand “nike” + a category “shoes” can generate “nike shoes”. While this is a simple example, our early users have used the feature extensively to guide their users to very precise content.
Guiding your user to the right content is paramount, and including a scope in the autocomplete is the perfect moment to do that. Simply indicate which attribute should be taken into account, and our Query Suggestions tool will generate the top related categories for each suggestion.
Easy to implement
After extracting and processing the data, the Query Suggestions tool will push all valid suggestions to a new index, and refresh it every 24 hours. Just like any other index, you can query it with the same API and libraries we provide. Here’s a tutorial with InstantSearch iOS.
Pricing & availability
Query Suggestions is available today for our Business and Enterprise customers. To enable this feature directly in your dashboard, reach out to your product specialist or customer success manager. All settings and blacklists are managed within the Algolia dashboard.