AI will change both “the why” and “the how” in the way web analytics is done
AI will change both “the why” and “the how” in the way web analytics is done. It will change both why we do web analytics and how we do it, Here is how.
Why do we do web analytic (the science)
The goal of web analytics is to make professional data-driven desisions on how to run your online business. The two core areas you look into is:
Traffic Management: The purpose is to maximise the return on investment on customer acquisition spending.
Content Management: Making sure the content support the sites business goals.
The purpose of Traffic Management is to maximise the return on investment on customer acquisition spending. It is done by using a “Learn and configure loop”. You learn what works and scale it. In other words a continuous improvement process.
In mayny ways, machines can do this better than humans because they can see patterns and adjust in real time. The patterns that the AI machines should look for are relatively easy to describe mathematically, and this makes this task ideal for a machine learning model. Canecto uses predictive analytics today to identify the best performing channels, and we can see that this area will evolve quickly when even more advanced AI models are applied.
The purpose of Content Management is making sure the content support the sites business goals. You focus on two areas Content and User Journeys.
Content: You focus on content is to ensure that the site is relevant for your target group so that they are motivated to do actions that support your business goals.
User Journeys: User journeys are the how users navigate on your site.
Today, Analytics on the last part is done using tools such as Funnels, A/B testing and via conduction of user surveys. However, all the methods have a joint weakness: They cannot be used as a foundation for predicting user behaviour because they are descriptive. In other words: they cannot be used as a base for AI or Machine Learning models. The approach has to be different.
Machines can do this better if you take another approach. You can train an AI model to understand the patterns and motivation of a user’s actions. We at Canecto has done this, and the results are great. We use Natual Language Processing and can come closer to understanding the intent of a user by monitoring how they interact with the content of a site. Such an AI approach can be used for making suggestions for both content improvement and user journeys.
The workflow of web analytics (the how)
Working with web analytics means extracting business value out of your analytics systems and applying your knowledge in two steps:
Data interpretation: Understanding the business consequences of the insights of the retrieved data
Notification: Reacting to findings with appropriate business actions
The current way web analytics is done is a classic IT-user system approach. You have a person how is “doing” web analytics.
There are two problems here, and both can to some extent be solved using AI. This first is timing and second is the ability to spot patterns and changes.
A machine can span and scan for pattern changes in real time, so they can detect significant changes in user behaviour more accurately and faster than a human. Canecto now has a set of notifications features that can do this for you.
A machine can also see complex patterns with more accuracy than a human can. It means that you will be able better understand how your users act on your site. At Canecto we use this principle to tell what was of interest for converting users. Such information lets the website owner know what content on a site tricked a specific action, and it makes it possible to make data-driven changes to a site’s content.
Web analytics and web content management will merge
The next logical step is to automate content publication based on user behaviour in real time. This will happen in two phases:
- Step 1: Content Management API: Support content and user flows
- Step 2: Deep integration into the publication of online content.
When this happens, then web analytics will be transformed into something else, because it becomes an applied science. It will be possible to have real-time website adaption of content and user flows – based on learned user behaviour.
Web analytics will be the engine that drives how a website performs. Both content and navigation will be personalised based on learned insights of predicted user behaviour.