Artificial intelligence and machine learning are two concepts that will lie on the lips of every marketer in 2018. Google is one of the parties involved pushing the research in this field to the next level. This year, we see shimmers of this technology coming through in Google Analytics. First there was Data Driven Attribution for GA360 users and now there’s Analytics Intelligence, a help module in Google Analytics that helps you find the right answer to your questions. The promises of artificial intelligence and machine learning are huge when embedded into analytics platforms. Not only do they provide better analysis, they’ll help democratize data analysis, opening it up to a larger audience. To see if Google Analytics Intelligence could live up to that expectation, we took it for a spin.
What is Google Analytics Intelligence?
Google Analytics Intelligence is a module built into Google Analytics that aims to answer basic questions on the user’s website. It does so by providing answers in the form of data and figures. The module allows users who don’t feel comfortable accessing the multitude of reports to find answers to their questions.
Now, where do you find the module? The module is accessible from every screen and can be found in the upper right corner in both GA and GA360. Do note that the feature is currently only available in English within accounts with languages set to English.
Clicking the Intelligence icon will open up a pane on the right hand side.
You’re now set to go.
Using Google Analytics Intelligence
As a user you have two options to kick it off:
1. Select a question from the dropdown. To do this, doubleclick on the text field. Suggestions will appear.
2. Phrase your own question. To do so, use the free text field. Note that not all questions will pass. The tool only does so depending on the linguistics and grammar you use, it also depends on the metric you add into the mix.
Capabilities and limitations
GA Intelligence will be using two of the most pertinent elements used in Google analytics, metrics and dimensions, as a basis for the questions received. Currently, questions can range from very basic to a level of somewhat higher complexity. We will briefly discuss the types of questions here.
Basic questions. Allows you to ask questions related to standard available metrics and dimensions in GA (e.g. ‘How many session did I receive on my website yesterday?’, ‘which channel did my traffic come from?’).
Performance questions. You can easily monitor your performances based on more specific metrics such as your conversion rates or landing page bounce rates etc. You could for example ask ‘Which channel converted the most for newsletter signups?’ or ‘which landing page saw the highest bounce rate?’ to spot those pages that might need review.
Interesting to note is that you can apply a filter by, for example, specifying a certain threshold (e.g. ‘which channel with more than 30 conversions saw the highest conversion rate?‘).
Trends. GA Intelligence will be able to hand you visual representations (read, graphs) of your data over time. Asking ‘How does organic traffic evolve over time?’ will for example result in a clear MoM overview of this traffic. If no time range is specified, GA Intelligence will apply a default last 30 day range.
Next, GA intelligence is capable of comparing data over different values (e.g. paid vs organic search) and over different time periods.
To have a better understanding of your data, you can ask for data in relation to the total (e.g. share of sessions coming from mobile?).
Lastly, the GA Intelligence module also supports questions related to e-commerce (e.g. product performance), advertisers (campaign, keyword performance etc.) and publishers (content performance, demographics etc.)
Of course there are still some limitations in the questions Analytics Intelligence will be able to support.
- General ‘how to questions’. The tool does not aim to be your new GA support center. Therefore, general ‘how to’ questions such as, ‘how to setup a custom dimension?’ will not be covered.
- Explanations. GA Intelligence will not be able to tell you why your data are behaving the way they do or why a certain trend occurred. However, Google claims to be already actively working on adding this.
- Strategic advice. The tool will not give any strategic or business advice in the sense of campaign optimization and strategy. Asking where to best allocate your investment will thus remain unanswered.
- General searches. Even though we are talking about a Google ‘search’ tool here, questions have to be Analytics related, asking for the weather forecast or a new cooking recipe will thus remain out of the picture.
With Google Analytics Intelligence, Google opens up the door of data analysis to the larger crowd. While Google Analytics Intelligence won’t directly replace your data analysts, it definitely provides a helpful tool for those who are newer to Google Analytics and are overwhelmed by the numerous reports and customizability the platform provides.
To wrap up, Google does also recommend you to test out the new feature and so do we. It will not only help you understand data better, but also benefit the learning of the tool itself which, in the end, results in more qualitative responses. This, of course, being the true beauty of machine learning and artificial intelligence.