The next big thing in analytics: understanding cause and effect in user behaviour

This is a guest blogpost by Adam Kinney: Head of Machine Learning and Automated Insights at Mixpanel

When it comes to data, machine learning (ML) is one of the hottest industry trends. But while ML is typically associated with process automation and IoT, there is a lot more that businesses can do with these insights. In fact, a growing number of organisations are using ML data to understand how new marketing campaigns and product features impact user behaviour.

It’s no doubt that machine learning has become more commonplace to help break through confusing, or even conflicting, observational data and give insights that can drive meaningful business impact. However, the problem is that user behaviour is very complex and does not necessarily follow pre-agreed rules. A machine can quickly evaluate the data but to successfully interpret this data and draw the right conclusions, businesses need advanced algorithms. This is why a lot of brands are starting to look at causal inference as a way to better understand user behaviour. Causal inference is a new trend within machine learning used to help marketers and business decision makers better understand the relationship between causes and impacts so they can make better decisions.

For instance, typically, people who frequently write product reviews buy more online than people who do not write reviews. If this is true and there is a causal relationship, it would make sense to encourage more reviews to increase revenues. But people who leave online reviews could also be a group of users who are more engaged with the brand than other users. This would explain why there is a correlation between willingness to buy and writing reviews, but not causal relationship. If that’s the case, then a marketing strategy encouraging customer loyalty will be more effective in driving customer engagement and sales than encouraging more users to leave reviews. Causal inference does exactly this – it assesses your current processes and allows you to zero in on the most important areas, so you focus your efforts in the right place.

It can also be used to help understand whether new product features or services are impacting user behaviour in the desired way. This could be done by controlling certain variables and analysing how they impact user behaviour. Of course, you can get answers like this from A/B tests as well, but A/B tests themselves take time and engineering work to run.

Causal inference allows businesses to simplify this process. The idea is to use statistical methods to create a model that predicts the most likely explanations for particular user behaviour. This simplifies decision-making processes and also allows companies to better allocate their data analysis resources.

Causal inference is only just beginning to move outside the world of academics into the business world, but I believe it will be the next big thing in Machine Learning and data analytics. Not only can it improve the understanding of how new product features and marketing initiatives impact user behaviour but, most importantly, it can enable businesses to innovate faster and stay ahead of the competition.