Ask and Get Insights on the fly with Google Analytics Intelligence

This post shares about a summary of what the Google Analytics Intelligence “Ask a Question” and “Automated Insights” features bring to the table for analysts and business users, how to use them, what some limitations are and how these features impact businesses today.
Photo: MarketingLand

We take a look at the latest Analytics Intelligence features released: Ask a Question and Automated Insights help users to ask for and receive “plain English” questions and answers via artificial intelligence.

Know what data you need and want it quickly? Just ask Google Analytics Intelligence and get your answer.

If you are new to using Google Analytics (GA), navigating the web platform and understanding all its terminology can be quite overwhelming. And if you’re not, here’s a truth about GA that many users might be able to relate to: unless you have deep familiarity with the tool or use it regularly enough, finding the information that you need in GA can often pose as a challenge. The sidebar menu on the left of the GA interface can be somewhat difficult to interpret or navigate for the occasional user, who may waste a lot of time poking around in report categories such as “Audience,” “Acquisition,” “Behavior” and “Conversions” to find the exact data that he would need for gathering insights.
These group of users, would then most probably approach the analyst team to get some help in getting the insights faster. A team at Google recently spoke to web analysts and indicated that they spend “half their time answering basic analytics questions for other people in their organization.” A report from Forrester found 57% of marketers find it difficult to give their stakeholders in different functions access to their data and insights.
Today, Google Analytics Intelligence latest “Ask a Question” and “Automated Insights” features aim to make it easier for people to be able to get the information that they require without needing to be well-versed with the Google Analytics interface or without having to be dependent on analysts. By asking a question in Analytics Intelligence in plain English, it is possible for business users and analysts to obtain analytics data easily and get access to the information that they need more rapidly, providing these both groups of users the ability to focus their energies on strategy and higher-value aspects of their roles.

What is Analytics Intelligence?

Analytics Intelligence is the umbrella term for a set of Google Analytics features such as Smart Lists, Smart Goals and Session Quality that use machine learning to help users understand and act on their data more efficiently on both the web interface and mobile versions (Android and iOS).

How do I use it?

On desktop, users will be able to see automated insights and type questions into a query box after clicking the “Intelligence” button in the GA toolbar on the top right hand corner.

Source: periscopix
By clicking on this button, it will open up the Analytics Intelligence side panel and here users can have access to the two new features of Analytics Intelligence, “Ask a Question” and “Automated Insights”. So what do they do?

Ask a Question Feature


Source: Google Analytics
Using Natural Language Processing, the “Ask a Question” feature in Analytics Intelligence provides an easy and clean interface to ask questions of your data. There is no longer a need for users to require deep background knowledge of GA reports to find out what they are looking for, as they can simply ask GA what they want to know and the relevant data will be pulled as answers to their queries. It can be likened to Siri or Google Now, within GA.
Ask a Question understands all standard GA dimensions and metrics. Additionally, it is able to pick up on the various naming conventions that are unique to your account. For example, it knows about all those Custom Dimensions and Event Labels you’re using.
However, one thing to note is that this feature is only available for English language users at the moment, but Google says it will add other languages as the system learns more about the types of questions users are interested in moving forward.

Source: WSOL Blog

Automated Insights Feature

The other main feature of Analytics Intelligence is “Automated Insights”, which will look through all your GA data and try to surface any trends, patterns or changes that may be hidden within reports.

Source: periscopix
In each of these insights, Analytics Intelligence will display the relevant data and provide a recommendation course of action to take. For example, these actions can be:
  • “To learn more, create a Segment with sessions that include: Device Category: mobile”
    (From the insight: Your site performs below average on mobile)
  • “Consider including these landing pages in your AdWords campaign strategy”
    (From the insight: Some high performing organic landing pages are not backed by AdWords campaigns)
The good thing is that you are able to save these Insights for later, and by doing so this insight will be pinned as a query to the Insights bar. Whenever you open your Analytics Intelligence, you can easily select this question to ask again. As these saved insights would be linked to your account, you do not have to worry about them being visible to any other users who have access to the GA property.
Source: periscopix
The goal behind these new features is to give the analysts and decision makers the information they want more quickly so that more time and resources can be spent on making use of the data for a positive change.

What kind of questions can I ask?

Currently, “Ask a question” supports typically “what” and “how many” types of questions about the data that GA knows about your business. These can include segments as well. For example, displaying sessions for the month segmented by age. Analytics Intelligence would then display the results with a chart, saving the user several clicks of effort and time spent digging.
Additionally, you can also ask follow-up questions based on the previous question you typed. There are suggested questions that display once a user clicks in the query box as prompts.

Source: WSOL Blog
Among the types of questions Google says users can ask Google Analytics:
What do you want to do?
You can try asking…
Basic reporting
“How many users did we get yesterday?” “Where is my traffic coming from?” “How many new users did we get last week on mobile?”
Check performance
“Which channel converted the best for [Goal X]?” “Which landing pages with over 500 sessions have the worst bounce rates?”
Chart trends
“Trend of new users this month?” “Graph of sessions from Chicago vs Seattle in December?” “Percent of Direct traffic over time?”
Compare data for different values or time ranges
“Conversion rate for referrals vs organic search?” “Average time on page for mobile vs desktop?” “How many [EventAction] did we have in February vs January?”
Ask about shares or percentages to understand significance
“Share of sessions by browser?” “What percent of sessions in the U.S. are from social?” “What share of sessions are from women?”
Ask complex questions combining multiple phrases
“How did share of new users compare in January for Firefox vs Chrome?” “Trend of new users this year vs last year”
Source: Google Analytics
Furthermore, they have several questions recommended for different groups or industry of users:
Which group of users do you belong to?
You can try asking…
For e-commerce websites
“Which products had over 200 unique purchases?” “What is our conversion rate in Spain?” “Share of revenue by country last quarter?”
For advertisers
“Top campaigns by [Goal X] conversion rate?”“Which banner ad content performs best?” “Which paid search keywords convert the best?”
For publishers
“Which [Custom Dimension for Article] got the most views yesterday?” “What share of people on [Page] are female?” “Trend of pageviews for [Author X]”
Source: Google Analytics

Limitations

At this time, Analytics Intelligence aims to answer questions about your historical Analytics data and is not currently able to answer questions related to the following:
  • General support (e.g. “How do I build a segment?”)
  • Explanations (e.g. “Why is my bounce rate increasing?”)
  • (Note: This limitation will be removed in the coming months.)
  • Strategic advice (e.g. “Which campaign should I invest in?”)
  • General search (e.g. “What’s the weather like?”)
Additionally, there will also be cases when Analytics Intelligence could not understand the terms or grammar used in your question and will be unable to answer.

What does this mean for businesses?

At present, Google Analytics have been already tracking massive amounts of data and generating reports for analysts and business decision makers. Now, this feature allows users across the organization to have the ability to surface insights in a quick-to-read format so they do not have to click around to navigate different pages to find these information, which make online business website performance tracking easy and saves time for more demanding or strategic tasks.
Specifically for larger companies that have a data analyst or team of analysts, these insights should help them scale up. And for smaller businesses that can’t afford their own analysts, this approach could bring that same type of possibility of having these analysts accessible to them and bring in insights that might be missed otherwise.
Does this mean that Google is killing the role of the analyst? While the Analytics Intelligence feature is impressive, one important factor to consider is that it cannot tell users what questions to ask in order to serve its business context best.
One of the major challenges that companies face today is that as the volume of data is increasing and more sophisticated tools made available for analyzing it, those data and tools are of minimal utility if the people using them don’t use the right metrics and ask the right questions.
Getting the most out of utilizing GA often requires a certain level of domain knowledge and experience that many organizations may not have and this may cause many in failing to maximize the value of this free tool. However, Analytics Intelligence may not fully be able to replace the role of an analyst as it is unable to tell users what they should be asking or what they should find out to be of real value to the business. In such cases, it is unable to suggest better and more relevant questions as well because it does not know what those questions are.
Furthermore, an example of a strategic question that Analytics Intelligence cannot answer is “Which campaign should I invest in?”. It is important for users to note that while the ease of using this new functionality provides a range of benefits, it will not help if it is used to produce reports and charts that do not actually help answer the key business questions at the end of the day.