View analytics data

Analytics give you insight into the usage trends, search quality, and end-user engagement of your app. The console provides an interactive dashboard experience powered by Looker. This page describes how to view analytics for your apps.

User events for analytics

This section describes the user events Agentspace uses to generate the search-related metrics, such as search count and click-through-rate. No additional configurations are needed if you are using the user interface to search and get answers through the Agentspace assistant. However, if you use the Agentspace APIs in your own apps, you must configure the search in your app to capture the user events and then import user events to your app..

Agentspace apps require the following user events to get the metrics:

  • Search events. Required for click-through-rate.
  • View-item events. Required for click-through-rate.

The following information is also used for some analytics:

  • Impressions. You can provide impressions (item views or clicks) that are attributable to Google by including the associated item in UserEvent.Documents when ingesting user events. This information is needed for attribution and to compute per-search metrics.
  • Attribution tokens. Attribution tokens are unique IDs generated by Google and returned with each search request. Make sure to include that attribution token as UserEvent.attributionToken with any user events resulting from a search. This is needed to identify if a search is served by the API. Only user events with a Google-generated attribution token are used to compute metrics.
  • Search logs. Search logs are based on search requests. To allow Google to identify user sessions so that it can deliver per-session analytics and higher quality search results, Google strongly recommends that you set the optional field SearchRequest.UserPseudoId.

If there are no search logs or user events, default values are shown for all metrics.

View Agentspace analytics

You can view search analytics for your app. You see the metrics related to the search calls made to your app. If your app is connected to multiple data stores, then making search calls to the app by specifying the engine resource is recommended so that you can get results from across all its connected data stores.

Use the following instructions to view analytics about your app.

  1. In the Google Cloud console, go to the Agentspace page.

    Agentspace

  2. Click the name of the app that you want to view analytics for.

  3. Click Analytics.

  4. Click a tab to view that metric group:

    • Adoption: Metrics on user adoption status.
    • Usage and Quality: Metrics on user feedback and app usage.
    • Agent: Metrics related to in-app agents, such as deep research, idea forge, and custom agents.
    • Value: Metrics on cost savings achieved by users querying data with the app.
    See the analytics data for adoption.
    See the analytics data for adoption

  5. To filter your metrics, specify the filters available for the metric tab and then click the Refresh button to apply them.

Metrics definitions

The following table describes how metrics are defined.

Metric group Metric name Metric definition Notes
Adoption Daily active users Count of active users per day A user is considered active if they log in and trigger any API calls, such as when landing on the home page.
Weekly active users Count of active users per week. The week starts on Monday and ends on Sunday.
Monthly active users Count of active users per month
7d retention rate Count of returned user in the current 7-day period / count of active user in the previous 7-day period A user is considered as a returned user if the user is active during the current 7-day period and the previous 7-day period.
7d growth rate The percentage difference between the number of active users in the current seven-day period and the previous seven-day period
7d churn rate Count of churned users in the current 7-day period / count of active user in the previous 7-day period A user is considered churned if the user was active in the previous 7-day period, but is not active in the current 7-day period.
28d retention rate Count of returned user in the current 28-day period / count of active user in the previous 28-day period
28d growth rate The percentage difference between the number of active users in the current 28-day period and the previous 28-day period
28d churn rate Count of churned users in the current 28-day period / count of active users in the previous 28-day period
Seats purchased The number of licenses (seats) the organization has purchased
Seats claimed The number of user licenses that have been assigned to and accepted by individuals within your organization
Usage and quality Search count Total number of search
Answer count Total number of query where Agentspace has generated an answer
Action count Total number of query answered by Agentspace Includes actions such as creating an email template.
Click-through rate (CTR) per search Total number of search clicks / search count Search clicks include clicks on search results and internal references.
Page view per category Total number of view for each category page
Feedback like and dislike count Count of likes and dislikes sent by users A record of the thumbs-up/thumbs-down feedback responses sent by the app's users.
Feedback dislike reasons breakdown Percentage of the dislike reasons When a user dislikes a generated answer, they can select multiple reasons to explain their dislike. The percentage shows how often each reason was selected.
Successful search rate Successful search count / search count
Successful answer rate Successful answer count / answer count
Agents Monthly active agent users Total number of active users who used the agent in the last calendar month.
Monthly agent chat sessions Total number of chat sessions initiated by the users interacting with a specific agent during the last calendar month. A chat session is defined as a user interacting with an agent within the same chat thread on a single day. If the user continues the conversation in the same chat thread on a subsequent day, it will be considered a new session.
Monthly agents used Total number of agents that had at least one active user during the last calendar month.
Monthly total chat sessions Total number of chat sessions initiated by the users interacting with the agents during the last calendar month. A chat session is defined as a user interacting with an agent within the same chat thread on a single day. If the user continues the conversation in the same chat thread on a subsequent day, it will be considered a new session.
Monthly agents created Total number of Agents that were created in the last calendar month.
Value
(public preview)
Successful search Count of successful searches. The number of successful searches are determined by the user behavior after a search. A search is considered successful if one of the following occurs:
  • If a search result is clicked before a new search is initiated.
  • If the response from a search includes a people search result, and isn't followed by another search within two minutes. This is because people-search results are displayed on the search result page and don't typically require a click.
Successful answer Count of successful answers An answer's success is determined by user behavior after it is generated. Success is measured by feedback provided or how soon a follow up question was asked.
  • If the user thumbs-up the answer, the query is considered successful.
  • If there is no feedback and the user does not ask another query within 30 seconds, the query is assumed to be successful.
Employee hours saved for selected dates (Successful search count + Successful answer count) x Minutes saved per successful answer
Value saved for selected dates Hours saved within the dates selected ✕ Average employee hourly rate To view this metric, in the Value tab, specify an hourly rate for the Employee Hourly Rate field.
Projected annual employee hours saved (Successful search count + Successful answer count) x Minutes saved per successful search The projected annual successful search count is derived from the successful search count within the chosen date range. Each query is estimated to save three minutes, an estimate based on Google's internal research.
Projected annual value saved Projected annual employee hours saved ✕ Average employee hourly rate To view this metric, in the Value tab, specify an hourly rate for the Employee Hourly Rate field.