Skip to main content

Implement a Google Analytics Use Case

A guide to implementing Auto Insights for a Google Analytics use case, including data structure and sample questions.

Auto Insights can help marketing and sales teams automate website performance analysis, drill into website visits, page views, sessions, or user trends as well as spot anomalies and outliers. However, it is important to note that the raw data from Google Analytics will require some data work to get it ready for Auto Insights.

This article will cover:

  • Example insights from this use case

  • Recommended data structure for this use case

What Sort of Insights Can Auto Insights Help Me Uncover?

We've outlined some example questions to provide an indication of the types of insights you can expect with Auto Insights:

Acquisition Performance Analysis

  • Visits by source this quarter

  • Year to Date trend in visits from mobile devices

  • Visitors by city this month

  • What’s driving trends in the number of bounces

  • Bounce rate by source

  • Number of entrances by page path

Behavior Trends and Analysis 

  • The average number of page views per visit

  • Average visit length by city

  • Average time on page-by-page path

  • Top page views by page path

  • Exit by page path

  • The average number of sessions per user

How Do I Structure My Data?

  • Auto Insights requires structured, transactional data, with at least 1 measure (e.g. Spend) and 5 segments (e.g. source of visit). In addition, we recommend at least 7 months of data (at monthly or daily granularity) so you can take full advantage of Auto Insights' Unexpected Changes feature.

  • For Google Analytics data it is likely that some additional steps to categorize data for Auto Insights' Natural Language Processing (NLP) will be required.

Example Data Structure

Please click on the picture to download a larger copy of the image.

Screen_Shot_2020-06-28_at_3_14_29_pm.png

Segments

Here are some of the typical segments we find in Google Analytics data. A segment is a qualitative value, like names or categories:

  • City

  • Device Category (desktop, mobile, tablet)

  • Host name

  • Page Path

  • Previous Page path

  • Source / Medium (e.g. referrals from sites like LinkedIn, or from email campaigns)

Measures

A measure is a quantitative, numeric value. Some of the typical measures include:

  • Bounces

  • Entrances

  • Exits

  • Page views

  • Sessions

  • Time on Page

  • Total Events

  • Users

  • Visitors (unique visitors)

  • Visits (number of visits)

  • Bounce Rate

  • Session length

  • Visit Length

Any Recommendations on the Process to Extract and Transform the Data?

Step 1: Extract

You may need to use a specific tool or solution to help you securely and reliably move data from Google Analytics into your databases, warehouse, or lake. Please reach out to us if you have any questions on this.

Step 2: Transform

Once you’ve extracted your data you may need to make some further transformational changes. This can either be done in your database itself or via creating an SQL query in Auto Insights. 

Nesta secção: