red line

Google's Analytics Intelligence Feature: Have We Reached the Singularity?

Google recently launched a new feature in its Analytics platform, simply referred to as "Intelligence." This feature brings the power of machine learning and artificial intelligence together to deliver actionable insights in the blink of a cybernetic ocular implant.

About Google’s Analytics Intelligence Feature

Ok, well Google may not be in a position to trigger a Skynet-level apocalypse (yet), but it’s AI tech is becoming pretty impressive. Recently, our team stumbled upon Google’s new “Intelligence” feature in Analytics, aimed at making insights more accessible to brands and marketers—and boy were we excited about the possibilities. From strategy to reporting, there are plenty of practical time-saving and valuable applications for this new feature.      

More is Less

You see, historically Google’s powerful Analytics platform has eluded or been severely underutilized by many brands and business owners. The sheer volume of data across dozens of specialized reports can make it very difficult to find meaning and distill raw information into a story that can be used to impact larger business objectives.

Putting it to Work

According to Google, the new Intelligence feature is pretty… intelligent.

How it works

Metrics and dimensions are the building blocks of reports, and you can ask about them in Analytics Intelligence using natural language. Learn more about how metrics and dimensions work. Let’s walk through an example:
  • Country is a dimension, and New Users is a metric.
  • In Analytics Intelligence, you can ask, “Which countries are most new users from?”
  • However, you don’t necessarily need to use the exact Analytics terminology when talking to Analytics Intelligence. You could also ask, “Which locations are new customers from?”. Analytics Intelligence will understand that “location” probably refers to the Country dimension, and “new customers” probably refers to New Users.
So is Analytics Intelligence the magic bullet brands and marketers need to draw actionable insights out of a sea of data?

Let the Testing Begin!

Scenario: Collecting Audience Data

We’ll keep it simple for starters. We asked Google to give us information about various audience attributes, including Demographics, Location, Device, and Frequency. Here are the results: Query: “visitor demographics” Google successfully pulled a report of Top Gender by Users—by default, reporting data from the last 7 days. By appending “year to date” at the end of the same search query, we saw the same report with the appropriately adjusted dates. Pretty cool! Query: “top locations over the past 12 months” By default, Google pulls the top countries, with the correct date range. By replacing the word “locations” with “cities,” we were presented with a deeper table-style report showing top metro areas (cities) by Session. Another one in the win column for Google.

Scenario: Collecting Acquisition Data

Often, marketing professionals and business owners want to develop a deeper understanding of where website traffic is coming from, so they can both build on any momentum their seeing, but also find ways to tailor content to specific audiences. For this next series of tests, we’ll dive into Google’s ability to tell us about incoming traffic. Query: “top channels over the last quarter” And we have a winner. Google pulled a Top Medium by Sessions report for the entirety of Q2, 2018. Time to turn the heat up. Query: “show me which pages saw the most traffic from Google last week” Let’s dissect this search query for a moment. In a nutshell, I’ve baked in a few different requests into this search. I’ve asked Google to fetch data on:
  1. Specific pages
  2. From a specific source
  3. Within a fixed span of time.
And Google wins again. The report it pulled was Top Page by Sessions for Source of “Google.” Awesome. Query: “which marketing channel is most effective?” Alright, we’ve reached a limitation, and for good reason. According to Google’s knowledge base quote above, Analytics Intelligence only understands explicit metrics and dimensions such as City, Medium, Conversion Rate, etc… In the example above, we saw very conveniently that “marketing” and “effective” are reported as terms that were not clear enough to produce a result.

The Verdict? Pretty Neat!

Google’s Analytics Intelligence feature may not be a perfect replacement for good old-fashioned data analysis, but I think it does a pretty good job at pulling simple-to-relatively-complex data in an instant.

Who’s It For?

This feature is tailor-made for entry to mid-level marketers, savvy business owners, and stakeholders in other departments who want to be more in-tune with how a client’s (or their own) website is performing. Overall, I expect more and more teams to take advantage of this tool in a variety of contexts—from reporting to internal strategy sessions and client meetings. I’ll certainly be encouraging some of our Marketing and Sales staff to give it a test run.
Related resources