Home < Technical < Product Overview < Autocomplete Analytics

Autocomplete Analytics

Prefixbox search box analytics provide an easy way to understand how your users interact with the search box and help you understand whether the configuration and data changes you are making to your search box improve the search experience or not.

The metrics are calculated daily and you can see the historical trends for each metric.

The following analytics dashboards are available in the Prefixbox Portal for all of your search boxes.

Search box overview

Generally speaking, we want to increase the suggestion usage and reduce the search box abandonment. Position 1 Usage gives an indication about the ranking quality of the suggestions. Our daily suggestion refresh mechanism helps you improve the suggestion usage by ranking more relevant suggestions higher automatically.

Coverage analytics help you understand how much your search box is able to show suggestions when users type and execute their searches. Conditional Usage tells you whether those suggestions were engaging to the users.

The daily suggestion list refresh helps in improving the coverage by adding more relevant suggestions to your suggestion lists based on your users’ previous searches.

Autocomplete suggestions help users save some of typing. We want to suggest the best suggestion even when the user did not type much (i.e.: minimize the Pattern Length). Intent Expression Time is a bit harder to reduce because it takes some time for users to read the suggestions, though it is important to track.

Search box audience

Increasing the Search User rate is important because users who search have higher engagement levels. Searching on e-commerce websites significantly increases the chance of purchasing a product. High quality search experience increases the Search User rate over time.

In order to deep dive into the above relative metrics you can monitor the absolute number of Searches and Search Users

Search box performance

You can monitor the performance of your search boxes near real time (with 5 minutes delay). It helps you understand how much load the search box gets and how fast it is able to serve the requests. If a search box is available in multiple data centres, you can check its performance in each data centre.