Machine learning in google analytics; what does this mean, how can it be used, and is this the start of the machine take over!By Alice Wheatley
Google have another new product to make our lives easier, and this time it is to help us to negotiate the data heavy minefield that is google analytics. Introducing Analytics Intelligence, the machine learning artificial intelligence data analyst imbedded into your google analytics.
But what can we do with this new piece of insight tech and how do we use it to pull insights into AMI campaign planning.
A key feature of google intelligence is that it learns as it goes, meaning the more our analysts use it (rate answers on relevancy and usefulness, search and respond to insights) the more accurate and useful the tool will become. Ultimately meaning you need a human touch to make this tech work to its full capability – the machines haven’t won just yet!
We have the ability to ask simple questions in plain English and get fast results. E.g. “Which channel had the highest goal conversion rate”, and Analytics will show us a ranked list of goal conversion rates by channel.
When we begin to engage with google intelligence the tech will learn from the team and begin to include insights on overall website changes and opportunities to be aware of, for example, it can point out that a certain landing page is performing better than normal, alerting us quickly to issues!
But what do these features mean to us, and how do AMI use these to add value to our life campaigns? Looking at a recent client example:
1. Google Intelligence told us there is a problem with page load speeds
2. Using the alert the average page load speed looks to be 11 seconds – this is high – try counting it out and thinking whether you would wait for a web page this long
3. Opening the page load report and changing the date range showed the insight team from start of the year to date, there has been a considerable increase over the year
4. Looking in more detail, highest page load time is over 70 seconds!
What (Human) insights can we take; page load times are a massive contributor to drop off, bounce rates and ultimately losing our well-earned leads before they even get to the website. At this point our team will do a deep dive into the website, unfortunately this cannot be done with MI or AI – yet! Using other pieces of tech (HotJar for example) we can pull together insights, why these drop offs are happening and if there are other website journey issues which we need to highlight. Ultimately making campaigns work harder and improving customers journey.
In conclusion MI & AI on google analytic is a great tool to alert our analytics teams to potential issues, then using their insights and expertise we can work on rectifying and monopolising on opportunities to make our campaigns work harder.