The Problem with Relying on Only One Predictive Analysis

By Anasse Bari, Mohamed Chaouchi, Tommy Jung

As you probably guessed, predictive analytics is not a one-size-fits-all activity — nor are its results once-and-for-all. For the technique to work correctly, you have to apply it again and again over time — so you’ll need an overall approach that fits your business well. The success of your predictive analytics project depends on multiple factors:

  • The nature of your data

  • The nature of your business and its culture

  • The availability of the in-house expertise

  • Access to appropriate analytical tools

The approach you choose will influence the model’s output, the process of analyzing its results, and the interpretation of its forecasts. And choosing an approach is no walk in the park. There are many things that can go wrong, many traps that you can fall into, and misleading paths you can take.

Happily, you can defend against these pitfalls by adopting a couple of wise practices early on:

  • Continuously test the results of your predictive analytics model. Don’t rely on the results of one single analysis; instead, run multiple analyses in parallel — and compare their outcome.

  • Run, test, compare, and evaluate multiple models and their outcomes. Use as many simulations as you can, and check as many permutations as you can. Some limitations in your data can only come to light when you compare the results you get from your model to those you get from other models. Then you can assess the impact of each model’s results vis-à-vis your business objectives.

Use multiple models to identify as many relevant patterns as possible in your data.