Predictive Analytics World in Boston has been a great show for the Experian Data Quality team. Data science is an emerging field and is just now hitting its stride. The things that companies can do with predictive modeling are incredible, and can be very valuable if done the right way.
Though I've been learning a lot since the event started earlier this week, there are three key things I wanted to share that are important to keep in mind when thinking about predictive analytics:
1. Focus on long-term thinking versus short-term gains. This simply means that your focus should be on long-term strategy, even when it’s unpopular. By considering where your customer will be five-to-seven years down the line, you can build towards the future versus sacrificing quality just for immediate returns.
2. Avoid over complication. Your focus should be on practicality and pragmatism. Just because a model or approach can be complicated doesn’t mean it needs to be. Identify the best model, solution or tool that supports your strategy, rather than picking a complex answer with lots of bells and whistles for no real reason.
3. Ensemble models can add real, incremental growth. Multiple models can be looked at like multiple players on a team adding different strengths and covering for other teammates’ blind spots. Consider including a number of models to help support your strategy.