Search analytics metrics
On this page
The Analytics page in the Algolia dashboard shows a variety of metrics that help you understand the performance of your search and how users interact with it.
Overview
The Overview tab shows at a glance how your search performs.
Total users
The Total Users chart shows the number of unique users who searched your site. By default, Algolia distinguishes users by their IP address. For a more accurate count, see Set the analytics user token. This is especially important if you’re searching from your backend.
Total searches
The Total Searches chart shows how often your users searched in the selected date range. As-you-type searches are aggregated if they occur within 30 seconds.
For example, the search “i” → “ip” → “ipa” → “ipad” counts as one search. The search: “i” → “ip” → “ipa” → 30 seconds → “ipad” counts as two searches: one for “ipa” and one for “ipad”.
Searches with the analytics parameter set to false are excluded.
For example, searches made in the Algolia dashboard set analytics to false by default,
so they’re excluded from your analytics.
For a detailed view of the popular searches, go to the Searches tab.
No results rate
The No Results Rate chart shows the percentage of searches with no results. For a detailed view of the searches without any results, go to the Searches without Results tab.
Click-through rate
The Click-through Rate chart shows the percentage of tracked searches with at least one click event.
A tracked search is a search with the clickAnalytics set to true.
If a single search generates multiple click events, it’s counted the same as if it generates one click event. Algolia doesn’t distinguish between different click event types and calculates one click-through rate for each record.
To calculate the click-through rate, you need to send click events.
Conversion rate
The Conversion Rate chart shows the percentage of tracked searches with at least one conversion event.
A tracked search is a search with the clickAnalytics set to true.
If a single search generates multiple conversion events, it’s counted the same as if it generates one conversion event. Algolia doesn’t distinguish between different types of conversions and calculates one conversion rate for each record.
To calculate the conversion rate, you need to send conversion events.
No clicks rate
The No Clicks Rate chart shows the percentage of unique searches that didn’t lead to at least one click event. It’s calculated as the number of unique searches without at least one click event divided by the number of unique searches. A unique search is the search after query aggregation.
Consider this example for (unique) searches for “iphone”, “ipad”, and “samsung”:
| Query | Day 1 | Day 2 |
|---|---|---|
| iphone | no click | click |
| ipad | click | no click |
| samsung | click | no click |
The no-clicks rate for the first day is 33% (1 out of 3 searches), for the second day 66% (2 out of 3 searches). Over both days, the no-clicks rate is 0% as all queries have been clicked at least once within the two-day time frame.
For a detailed view of the searches without any clicks, see the Searches without Clicks tab.
To calculate the no-clicks rate, you need to send click events.
The click-through rate and no-clicks rate don’t add up to 100% of searches.
Click position
The Click Position distribution shows the position of a search result with clicks. Smaller values are better since they indicate that users clicked top results.
Consider the following click events for the query “ipad”:
- First result: 2 clicks
- Third result: 1 click
- Tenth result: 1 click
The average click position is (1 + 1 + 3 + 10) / 4 = 3.75.
To calculate the click position, you need to send click events.
Searches
The Searches tab lists your most popular search terms. For each search term, you can check the following metrics:
| Metric | Description |
|---|---|
| Query | The most popular query terms |
| Count | The number of times they’ve been searched |
| CTR | Click-through rate |
| CVR | Conversion rate |
| Click Pos. | The average click position |
| Total Searches | The percentage of all searches made up by this query term |
Select a search term to see more details:
- The results showing for this search term
- The Average click position and a click position histogram
- Any Rules or filters applied to this search
Popular search data helps you examine your product vocabulary. You can align your users’ vocabulary with that of your database. For example, if users search “sofa” more frequently than “couch”, your database should support both “sofa” and “couch”. You can do this by adding attributes or synonyms.
One of your most popular searches might be <empty query>,
which is a search request without a query.
If you want to exclude it from your analytics, see Exclude queries.
Searches without results
The Searches without Results tab lists all searches with no results. For each search term, check the following metrics:
| Metric | Description |
|---|---|
| Query | The most popular query terms without results |
| Count | The number of times they’ve been searched |
| Total Searches | The percentage of all searches made up by this query term |
Select a search term to see if any of your Rules or filters were applied.
Searches without results can have two reasons. Either your users are searching for something you don’t offer, or they can’t find it. Consider expanding your catalog if there’s a genuine gap in your index. If users can’t find what they’re searching for, you can either add the right keywords to the relevant attributes, or add synonyms or Rules.
If you add a rule to show a banner for a particular search term without results, you still see it in your Searches without Results since it isn’t retrieving any records from your index.
Searches without clicks
The Searches without Clicks tab lists all searches with a 0% click-through rate. For each search term, you can check the following metrics:
| Metric | Description |
|---|---|
| Query | The most popular search terms with 0% click-through rate |
| Count | The number of times they’ve been searched |
| Total Searches | The percentage of all searches made up by this query term |
Select a search term to see more details:
- The results showing for this search term
- Any Rules or filters
Searches without clicks are underperforming queries. You might add synonyms or Rules to boost results that your users would find more appealing.
To see the searches without clicks, you need to send click events.
Results
The Results tab lists the most popular search results. For each result, you can check the following metrics:
| Metric | Description |
|---|---|
| Results | The most popular results across all searches |
| Count | The number of times they’re returned |
| CTR | Click-through rate |
| CVR | Conversion rate |
Filters
The Filters tab shows your most popular filter attributes and values. You can check:
| Metric | Description |
|---|---|
| Top filtering attributes | The most used filter attributes—for example, “color”, “category”, or “brand” |
| Popular values per filter | The most used filter values—for example, “red”, “books”, or “apple” |
You can use these analytics to improve your faceting configuration. For example, if your users filter on “size” and “color”, but not on “brand”, remove your “brand” filter or display it lower on your page.
The same applies to filter values. If users filter the colors red and yellow more often than purple, you can adjust your inventory by stocking fewer purple items, or adjust your index settings by promoting red and yellow over purple.
Geo
The Geo tab shows the geographical origin of the searches and the associated network latency.
Location metrics help you learn who your users are and where they come from. Understanding latency can help you better choose where to place your servers and DSNs.
