Looker Connect Training
Help Center
Documentation
Community
Cloud Certifications
We built a Sessions table derived from Events, that tracks only the 10 most important events that can happen during a session. We had 200+ events for which I could not see their impact on Sessions conversion rate and performance: this would have been too costly in terms of memory. The Sessions derived table also included very little Page URL details: too costly to include tracking for the performance of each different banner ad referring traffic. We could only access that level of detail at Events level. The standard pattern in our main Sessions table, for each key event: SUM(CASE WHEN event = ‘MainFunnel’ … ) etc. Users were asking me “How might we see the impact of this specific Tooltip?” “What was the impact of version x.1.2. of the ABTest?” “How did users who came through this tweet convert compared to that post?” “What about users on this new OnePlus?”… Anything so granular could not be addressed in their standard Sessions Explore. The syntax below allowed me to let a User pick t
How to allow end users to compare their custom date ranges in the Explores in just one query? Comparing 2 or more standard date intervals such as week, month is straightforward thanks to Looker’s time filters. However our teams may also require custom date ranges. We release a new feature or launch a new marketing campaign at irregular intervals. We want to confirm their impact as soon as possible, and we keep track daily of the statistical significance of variations observed. This may result in comparing the 11 days before the change with the 18 days following. In Looker an end user with no access to Developer mode can easily apply some types of filters: filter 1 period through a Date range filter. compare several values a dimension can take to the rest of population: after Data team developers have created the filter and dimension as specified by the Looker team here, any end user can compare measures for 2 values of one dimension, eg ‘Brand’, in one step. They simply select ‘Brand’
In a Visitors table, I had created a dimension for the Visitors’ Time to Order, as well as Time to complete several interactions. The Visitors table is derived from Sessions by grouping them by user_id. Time To Order was of course useful, but I found it would be great to separate those users who purchase on 2nd visit, perhaps one month after their first session, and those users who after their first session make n visits during the first month, but only eventually purchase at the same time as the other group. My code for this is along these lines: derived_table: SELECT s.visitor_id , MIN(CASE WHEN s.order_count > 0 THEN s.run_sequence ELSE 100000 END) AS run_sequence_of_session_with_first_order FROM sessions AS s GROUP BY 1 fields: - dimension: run_sequence_of_session_with_first_order label: '._. CONVERSION - Sessions To Order: session number ... has 1st order' type: int sql: | CASE WHEN ${TABLE}.run_sequence_of
Already have an account? Login
No account yet? Create an account
Enter your username or e-mail address. We'll send you an e-mail with instructions to reset your password.
Sorry, we're still checking this file's contents to make sure it's safe to download. Please try again in a few minutes.
Sorry, our virus scanner detected that this file isn't safe to download.
