Solr Search Analytics Insights - SearchStax
SearchStax® by Measured Search® includes search analytics reports that track the queries and search results of your Solr-as-a-service deployment.
To view the displays on this page, you must:
- Create an analytics app.
- Insert analytics scripts in your search-results page.
- Let people use your Solr application so you can collect statistics.
- Navigate from the Analytics node on the left-side navigation column to the list of Analytics Apps.
- Click the blue arrow icon to the right of the appropriate App.
- Select the Insights node in the navigation bar at the left edge of the SearchStax window.
Sources of Insight
The graphs shown on this page are based on the searches submitted to your application over a specific time period: this week, last week, this month, or last month.
Imagine a list of query keywords ordered by frequency (the number of queries that they appeared in). The most-popular keywords might be:
This is the same as the list of Top Searches.
Now take this list and turn it on its side, with the most-popular keywords at the left. This is essentially what you see in the Unique Searches Distribution graph, below on this page.
The Search Query Rank graphs use this list of keywords as the X-axis of a graph depicting some associated attribute on the Y-axis. For instance, how does the frequency of no-result searches change as we go from most-common keywords (on the left) to most-unique keywords (on the right)?
The purpose of these displays is to help you optimize the most-popular queries without being misled by the kind of random noise that the least-popular queries can create. The least-popular (most unique) queries often include nonsense words and misspellings that have no relevance to your application.
Search Query Rank
|MRR||Mean reciprocal rank.|
|Average Click Position||The average position of the result item when this query resulted in a clickthrough event.|
|No-Result Searches||Percent of searches that produces zero results.|
On these graphs, each bar of the graph represents the cumulative total of the top N% of the searches. For instance, "10%" means the keywords associated with the top 10% of searches. However, "20%" means 0-20% of all searches. "30%" means 0-30% of all searches. "100%" means all searches.
Unique Search Distribution
This graph ranks search keywords by the number of searches they were used in. In this example, just three keywords accounted for 10% of the searches during the selected time interval. (Many people searched for those three terms.) Ten more keywords accounted for the next 10% of all searches. Click on a blue number to see a table of the keywords and their frequencies.