Causal inference is a new trend within machine learning used to help marketers and business decision makers better understand causes and impacts so they can make better decisions. If this sounds like a foreign language to you, then you’re not the only one. Causal inference is only just beginning to move outside the world of academics and research scientists to what it is today – a more relevant asset for businesses, so it’s time to get well acquainted with the language.
Throughout my career, every team I’ve worked across different brands has incorporated machine learning differently. With so much information being collected by companies, one challenge they face is how to process the large datasets of information in order to unlock deeper insights. It’s no doubt then 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.
When it comes to machine learning, in my opinion, there are two key scenarios where companies can use causal inference to help with innovation and growth. Firstly it can be used to help with the planning side of setting goals or objectives and secondly it can be used to help understand the impact of new features.
No more trial and error
Most companies will have high-level objectives in place that align with business goals. These can be growing the user base, reducing customer churn, or increasing conversions. But it’s difficult to know what changes are required for your product or marketing to achieve these objectives. Typically, companies will try out different approaches and see what works best. This trial and error method of testing is expensive and time-consuming. Each experiment involves development time and/or marketing spend. It would be much more efficient to learn from the data you have already collected to see which areas of your product or marketing will likely get you to your goals. 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.
Evaluating the impact of your actions
After implementing a new feature, you need to assess whether it has got you closer to achieving those goals. This isn’t always an easy process as high-level goals tend to be very difficult to move in dramatic ways. You’re more likely to succeed with small, incremental increases over a period of time. But each of those small incremental increases can get lost in the noise of day-to-day fluctuations in your KPIs. This is where causal inference can offer significant benefits in helping you understand if the newly launched feature causes users to behave in a way that will get you closer to your goal. For example, you can see if that new email digest causes users to churn less. 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 for many product features. Here are some examples of causal inference in action to help streamline processes and ultimately meet business goals:
In this example, a digital media site that has a subscription paywall wants to reduce the number of consumers that cancel their subscriptions. To meet this objective, the site can start off by comparing users who have cancelled their subscriptions with those who haven’t. Results show that users who read at least three stories a day on the site are those who cancel less often than users who read less than three stories. Does that mean that if you get more users to read at least three stories per day cancellations will decrease? Not necessarily. It could be the case, for example, that the users who are most passionate about the media site like to read a lot of stories and are less likely to cancel than other users. In that case, reading more stories doesn’t cause users to be less likely to cancel, even though reading more stories is correlated with being less likely to cancel. In this scenario casual inference can help you understand which top three co-relations are the most likely reasons for the customer behaviour of the users who choose to cancel their subscription and make the right choice about how to reduce cancellations.
Let’s say you also find that users who sign up for the media site’s email newsletter cancel less often than those who don’t. This doesn’t necessarily mean that there’s a direct correlation here, just that users are both signing up for the newsletter and keeping their subscriptions. On the other hand, maybe the newsletter is a daily reminder for users of the value they get from their subscription, and it causes users to be less likely to cancel. In this case, getting more users to sign up for the newsletter is directly related to getting less cancellations.
Causal inference is a technique that allows businesses to identify whether the causal explanation is true in scenarios like those outlined above. This works by controlling confounding factors. In the first example of users reading three stories a day, you can control the super-user confounder by predicting which users are most likely and least likely to read three stories per day (we call this prediction of the user’s propensity to read three stories per day). And then seeing if there is a difference in cancellations between the users that really do read three stories per day and those who don’t, within each propensity group.
As causal inference becomes increasingly popular for businesses, it’s important to be able to tap into tools that are already using it in order to help your teams innovate and grow more quickly to stay ahead of the competition.
Adam Kinney, Head of Machine Learning and Automated Analysis, Mixpanel