Click and Conversion Analytics

Click and Conversion Analytics takes Algolia’s out-of-the-box Search Analytics further by providing insights into actions users take after performing a search. Click and Conversion Analytics also form the basis for more advanced features like A/B testing, Dynamic Re-ranking, and Personalization.

By default, Algolia doesn’t track what happens after returning search results. It doesn’t know what you do with the results, nor what your users do with them. For Algolia to calculate Click and Conversion Analytics, you need to send information in the form of events whenever your users click on search results or take meaningful actions like bookmarking, sharing, or purchasing items.

Algolia uses click and conversion events after searches to compute metrics such as click-through rate (CTR), conversion rate, average click position, and other analytics metrics of your searches.

You can also send click and conversion information for your users’ actions outside the search context. These won’t contribute to metrics like CTR or average click position, but Algolia can use them for Personalization.

Click and Conversion Analytics let you measure the quality of your search relevance. For example, you can see which queries have consistently high CTR or conversion rates and which items your users think should be higher in the ranking. You can use this information to tweak your relevance settings to push these items higher in the search results. Dynamic Re-ranking automatically promotes items based on your CTR and conversion rates.

What are click and conversion events?

Whenever a user clicks on a search result or clicks on items that reveal their affinities, you can send Algolia a click event. If they go on to add a search result to their shopping cart or purchase that item, you can send Algolia a conversion event. You define what your conversion events are, whether purchasing an item, reading an article, listening to a podcast, or something else.

You send events by adding small bits of code to your application. For more information on which events to collect and what information to include in an event, read the guide on capturing user behavior as events.

Click and Conversion Analytics correspond to user actions and the related events after an Algolia search. A/B testing and Dynamic Re-ranking use these analytics. However, you can use these events for more than Click and Conversion Analytics.

Click and conversion events outside the search context are helpful in Personalization purposes. Even if you aren’t planning on implementing Personalization immediately, it’s best to consider sending events for the eventual possibility that you will. That way, you won’t have to reformat your events for Personalization, potentially losing out on previously gathered data.

Implementation steps

Implementing Click and Conversion Analytics is a two-step process.

1. Plan which user actions to gather.
2. Capture user actions by sending events.

Before sending events, you should first plan which user actions to gather. Once you’ve made a plan, you can send events to Algolia using the front-end libraries and API clients. A REST API, called the Insights API, lies beneath these clients and has a single endpoint. For this reason, Algolia sometimes refers to these events as “Insights events”. Because different event types require specific information, it’s best to use the InstantSearch libraries or API clients, which provide distinct methods for sending different event types.

Once you’ve sent events, you can validate that you’re capturing all the necessary information using the Events Debugger.

Insights events (view, click, and conversion) don’t take immediate effect. The delay can range from a few minutes to an hour, depending on whether they’re sent after a search or not, and how long after a search they’re sent. For a better estimation, see: How long does it take for Insights events to be taken into account?.

Using Click and Conversion Analytics

You can access Click and Conversion Analytics either in the Algolia dashboard or programmatically via the Analytics REST API. To show how these metrics can help, consider the following three queries that use different words to convey the same intent: to find an eco-friendly refrigerator.

Click-through-rate

Suppose the click-through rate for the three queries is as follows:

• “refrigerator efficient” - 50% CTR
• “eco refrigerator” - 10% CTR
• “refrigerator saves energy” - no results, therefore 0% CTR

Using this information, you could take the following actions:

• Add synonyms between “eco refrigerator” and “refrigerator efficient”.
• Creating a three-way synonym with “eco” and “efficient” and “saves energy”
• Index additional information to your products to use these words in all your records.
• Create a filter or a category for “eco-friendly”.
• Use the word “efficient” more often in your product descriptions, or promote products that describe themselves as “efficient” since your users seem to prefer the word “efficient” over the word “eco” (50% versus 10%).

Conversion rate

Suppose, after making one or more of the previous changes, the conversion rate for the three queries is as follows:

• “refrigerator efficient” - 10% conversion rate
• “eco refrigerator” - 5% conversion rate
• “refrigerator saves energy” - 5% conversion rate

These are some insights you can glean:

• Including these terms in the indexed data was a good move because they all convert.
• Given the higher conversion rate on the “efficient” query, it seems to be the preferred term.
• If you see lower click or conversion rates than you would expect, it could be that none of your refrigerators are eco-friendly. You might consider adding these items to your product line if this is a popular search.
• If you don’t have enough of the highest-converting items in your inventory, you might want to improve your stocks.