Predictive analytics, at the most basic level, is about having actionable data about what customers do in real-time via their phones. It’s actually somewhat amazing to consider how far we’ve come in this regard.

Consider a marketing team in 2000, in an era we’ll call “Pre-Predictive Analytics.” (Honestly, many marketing teams are still in this era, but we’ll gloss that over for now.) In order to get the best customer insights possible, a marketing team would (in collaboration with other departments) need to create dozens (hundreds?) of spreadsheets and then pore over them for insights and patterns.

Now flash forward 16-17 years. Marketing teams who have embraced mobile can use machine-driven technology to target specific customers via their phone, which is almost universally in their pocket. One of the golden rules of sales has always been that the more you know about a lead, the better your chance to convert.

If we were going to sum up predictive analytics in 3 steps, it would look something like this:

1. Know a lot about a lead;

2. Tailor content to that lead;

3. Adjust the end user experience for that lead, not universally.

It’s incredibly powerful for marketers.

Here’s an example: OneFootball, who works with us, wanted to identify users with the highest risk of churning (turning over) — and then proactively target them with individualized messages during the football (soccer to U.S. fans) offseason. They used a predictive analytics platform and reduced churn by 7.5% in the offseason. EyeEm, similarly, reactivated 67.3% of customers at high risk of churn with predictive analytics.


Predictive analytics can also inform product decisions. Some U.S.-based companies design their mobile app with a U.S. audience in mind, then find (via predictive analytics) that another market churns at a huge rate because the app isn’t geared at them. In such a situation, the company could make a Europe-focused app or Asia-focused app and reduce churn in those markets, which could lead to revenue uptick.

A 2015 Juniper study on digital retail marketing showed that the market for mobile predictive analytics might approach $200 billion in the next few years, largely because of the effectiveness in targeting ads to specific users.

The market for mobile predictive analytics might approach $200 billion in the next few years.

Another benefit of predictive analytics is connecting different online, on-mobile behaviors to offline, in-store purchases. What aspects of a product can you feed someone on their phone that will drive them to come into a store and make a purchase? It varies by specific product, but US Cellular used predictive analytics on mobile engagement vs. in-store purchases and found, for example, that offers, device galleries, and plan galleries were the most effective. Site search, compare phones, and the mobile support URL barely registered. This can influence future app design, future responsive site design, and other decisions around marketing and sales.

You can even blend mobile predictive analytics — which some might call “new school marketing” with direct mail, which is generally fairly “old school marketing.” Some companies use their mobile prediction suites to determine which customers will respond favorably to snail direct mail, for example. It gives them another point of attack in the marketing mix.

Predictive analytics can also benefit you from an organizational standpoint, helping to see beyond silos. What does that mean? In most conventional business set-ups, each silo (department) takes customer data and sorts/analyzes it by their specific need — perhaps even projecting into the future by their needs. With predictive analytics, one set of user behavior can be the baseline data set that everyone operates from; analytics can then drive decision-making, as opposed to potentially being used as a source of disagreement among decision-makers.

Marketing has changed a tremendous amount in even the last decade. Most business school students going into the field are still being taught The 4 Ps, and while they remain relevant, a better model might be 5S: science, storytelling, speed, simplicity, and substance. Predictive analytics is definitely scientific, allows you to tell a personalized story, works quickly, is fairly simple, and is the essence of substance (an offer the user wants, not one you’re pushing to everyone). It might be the face of new-school marketing, as a result.