Predictive Apps vs. Predictive Analytics. Is it really a battle between the two or is one a subset of the other? Many of us use what I would describe as predictive apps, in our consumer life, day to day but does the idea extend beyond the consumer realm and into the business? How does predictive analytics fit into the picture?

Below are a few of my thoughts but I am also interested in what you see as the difference, if any. Would you use a different phrase?

Setting the scene

It’s time. Time in my opinion for businesses to match the utilization of analytically driven, automated processes offered by the consumer apps we are using day to day on the web or our mobile devices. Google, Amazon, Facebook, Twitter and more offer consumer focused predictive apps with embedded and automated processes driven by predictive analytics. These predictive apps provide the rich and tailored experience I have grown to expect. Teams of data scientists have worked their magic with predictive analytics which then are embedded to drive automated decision making, in processes, which ultimately deliver the intimate and guided experience which delights me and draws me further into their app and further into their world.

“It is not predictive apps vs. predictive analytics. Predictive apps embed predictive analytics in a context aware manner to drive automation and a tailored experience.”

So if we extend the quote above then the second you embed predictive analytics into a specific, and reusable, business process you essentially create a type of predictive app. For example consider Amazon on the web. When you browse items you get a list of “things others who looked at this item have looked at” and when you select an item to buy you get a list of “things others have bought who bought this item”. The predictive analytics component in the background is a recommendation engine which continuously and automatically learns using historical and current data from ALL users. The business process is your regular e-commerce buying process which has been automated to try to drive cross-sell through the use of predictive analytics using information you are providing whilst interacting. Most importantly the things being shown are tailored to you as an individual based on what you are doing right now, not someone that looks like you, and what you and others have previously done.  Could this be an example of a consumer focused predictive app which adjusts in real-time, automatically through the automatic application of predictive analytics or is it just an analytics enabled predictive analytics business process?

Another way to think of this is that when you combine business process understanding, predictive and self-learning analytics, historical/current data and analytic driven automation you have a predictive app or at the very least a predictive analytics driven business process.

Self-contained

Theoretically a predictive app should be able to be seen as self-contained. What I mean by this is that you know you have a predictive app when you could open it up to third parties by providing a standard input and standard output interface. Initial input data in, combined with historical and ongoing additional data inputs, and predictive analytic driven results out. In this case self-learning predictive analytics is used to add intelligence and drive the process automation throughout.

The next level

Now take it to the next level. Imagine that in any one enterprise business process there are 20 sub-processes. Essentially you could look at each sub-process and work to build them into predictive apps by adding predictive analytics to facilitate automation. From there you could orchestrate them together and deliver a larger predictive app which could power an entire part of the business. This potential layering is important. It is what will enable organizations to take parts of their existing legacy processes and to slowly introduce predictive analytics transforming parts of them into predictive apps without having to replace everything. It could also be the engine to monetize  services through open APIs.

Your thoughts?

I published this question on Twitter to my followers. Below are a few of the responses.

I would love to hear yours as well (click here to get to the tweet and reply), or leave a comment below. I would love to find out if you agree with these responses or the position taken above.

In Conclusion

In my opinion it is not a battle between predictive apps and predictive analytics. The real question is one of if you want to buy a black box predictive app and have it maintained for you, if you want to develop your own predictive apps (or analytically driven business processes) by adjusting existing processes and introducing predictive analytics to drive automation or if you want to continue along safe in the knowledge your competitors are going to do it more intelligently. The battle is to decide to change and then working out the most effective way to do it!