Analytics

Get to know your Search.

Analytics and data play a crucial part in almost everything; you can use the analytics section in the Sigmie application to see metrics about your Searches and your users.

Engine

Unlike the traditional request counting that's used for counting requests made to your Sigmie application. There is a difference when counting Analytics Requests. Each time you user types a letter in a search field a new request is made. Counting those requests would lead to confusing data in your analytics.

The analytics engine sets a search session duration of 30 seconds. As long as the user types in the Search box and the last input matches the previous one this is counted as a single analytics request.

In the example

• p -> pe -> pet -> pete -> peter the user types 5 letters making 5 requests but they are counted as 1 in the analytics engine.

• peter

Now in the example

• p -> pe -> pet -> m -> mi -> mic -> mick -> micke -> mickey the user makes 3 requests for pet and then he changes it's mind and starts seaching for mickey making in total 9 requests, 3 for the letters pet and 5 for mickey. These are considered 2 different searches and are stored as 2 analytics reuqests.

• pet
• mickey

In the last example

• p -> pe -> pet -> (30 seconds passed) -> pete -> peter The user types 5 letter making 5 request but between pet and pete passed 30 seconds. In this case this search is treated as 2 analytics request.

• pet
• peter

Users

The analytics engines distiguishes the requests made by the IP Address, but if you can also send a user_token to use instead.

Example

{
  "query": "peter",
  "user_token": 34
}

Metrics

In the Analytics section in the Sigmie UI, you will find various metrics about your application Searches. Each metric also has a trend indicator that compares the values of the current time range with those of the same previous range.

To clarify, if your time range is "Last Month", this trend compares the retrieved values with the "Month before the last".

Total Searches

The "Total Searches" metric shows how many Search requests were made during the specified time range.

No Results Rate

The "No Results Rate" displays what percent of the Searches didn't return any matches.

Unique Searches

The "Unique Searches" displays how many incoming Queries are unique.

Response time

You Index and Search settings can affect your search engine's response time.

This metric shows the average time your engine needs to process the incoming Query.

Total Hits Served

"Total Hits Served" is the total amount of the returned hits.

Search Requests The "Search Requests" graph displays the Search requests spread over the time range.

Scoreboards

Popular Queries

The "Popular Queries" show the Queries that your users typed the most. This is useful to understand "how" your users use the Searchbox that you gave them to make adjustments accordingly.

Top Queries without results

The "Top Queries without results" are the Queries that the users type more often, and no results are returned.

For an Online-Shop this would indicate a product that may be missing from your Shop.

Suppose the word "Nutella" appears there. In that case, it means many users come to your Shop wanting to order a "Nutella", but it isn't available.