Data dictionary: VA.gov Search KPIs
Definitions of metrics used on the Standardized VA.gov Search KPI dashboard and guidance for how to interpret them.
About search products
The search_scope filter at the top of the dashboard can be used to filter by search product. Search products are defined in GTM and configured by pulling the 3rd item from the eventAction field in Google Analytics.
The one exception is the “VA Homepage” product, which is a subset of the “All VA.gov” product. This data only includes searches that originate on the VA homepage or results clicked immediately afterwards. If the user searches again instead of clicking a result, the initial search will be counted as a refinement, but the second search will not be included in the “VA Homepage” data, as it originated on the search results page instead.
Highlights
Chart title | Definition | How should I use this? |
|---|---|---|
Successful searches | Total number of clicks on search results. | An increase means more people are clicking on results! A decrease could be fewer people using the search or that fewer users are finding results they want to click on. |
Search funnel | Searches: Total searches performed. Clicks: Total clicks on search results. | The closer these numbers are, the more users are clicking for every search they do. The further apart they are, the more users are searching without clicking any of the results they receive. |
Search refinement rate | Total search events where the user entered a new search immediately after seeing results, instead of clicking one of the titles ÷ total search events. | A high refinement rate means that users aren’t finding what they need and have to try their search again instead of clicking on results. |
Search abandonment
Chart title | Definition | How should I use this? |
|---|---|---|
Search sessions | Total sessions in which a search was performed. | Changes in total number of search sessions may impact search abandonments, as well as successful searches. Read these KPIs against each other to better understand whether you’re seeing changes in search usage vs. search outcomes. |
Abandoned searches | Total sessions where a user searched, but did not click any results. | This metric could be impacted by changes in overall searches or by changes in how many users saw results that were or weren’t useful to them. |
Search abandonment rate | Total sessions where user searched but did not click a result ÷ total sessions where user performed a search. | Of sessions where users searched, how often did they not click on any results? Increases to this rate may indicate that fewer users are seeing helpful results in their searches. |
Abandonment rates by most common search terms | Search terms: Text the user entered for their search. Searches: Total number of search events. Abandonment: Total sessions where user searched but did not click a result ÷ total sessions where user performed a search. | By looking at the most frequently searched terms and highlighting the ones with higher abandonment rates, you may be able to identify popular search terms where results are not what the user expected. |
Search refinement
Chart title | Definition | How should I use this? |
|---|---|---|
Refinement rate | Percent of searches where the user entered a new search after seeing results, instead of clicking one of the options presented. | A high refinement rate means that users aren’t finding what they need and have to try their search again instead of clicking on results. While the search refinement distribution above shows how many times users repeatedly refined their searches, the refinement rate shows the percent of searches where any refinement occurred at all. |
Search sessions by number of refinements | The distribution of how many times each user entered a new search after seeing the results instead of clicking one of the titles. Count of search events is grouped by session. | If the search is returning more helpful results, we’d expect to see the distribution of sessions shift to the left (more sessions with fewer search refinements). |
No results found | Total number of times when no results were found ÷ total search events. | This may be driven by changes in search performance or in what topics users are searching for. |
Search completion
Chart title | Definition | How should I use this? |
|---|---|---|
Search click rate | Total clicks on search results ÷ total search events. | What percentage of all searches did users click on a result? In an ideal world, this would be as close to 1 as possible! |
Clicks by position | Total clicks on search results, broken out by the position of the link in the results returned. | Higher click rates for 1st or 2nd position results mean that the user chose to click on the top options the search returned. When users click on results further down the page, it indicates that the top results returned by the search were not useful to them. |
Results clicked | Results chosen: Titles of the most frequently clicked search results. Total clicks: Number times that result was clicked on. | The top results chosen show you which links your users are clicking most often. This can provide helpful context for what your users are looking for (and what they’re finding). |
Finding search products
Chart title | Definition | How should I use this? |
|---|---|---|
Top pages viewed before searching | Page location: most common parts of the VA.gov website viewed before search was performed. Views: Total searches performed. | Where are your users coming from directly before searching? Does it match your expectations? |
Total search users | Unique visitors who performed a search. | This should provide context for other KPIs on this dashboard, which may be impacted by changes in overall search use. |
Avg. searches per user | Total searches performed ÷ users who searched | This should provide context for other KPIs on this dashboard, which may be impacted by changes in overall search use. |
Search users by device | Users who performed a search, broken out by category of device used. | How are people accessing your search product? This information can be used to identify potential device-specific bugs or prioritize future improvements. |
Top browser types used to search | Users who performed a search, broken out by type of browser used. | See above. |
Help and feedback
Get help from the Platform Support Team in Slack.
Submit a feature idea to the Platform.