How Not To Sabotage Your Predictive Analytics Project
Most of us are familiar with many of Peter Drucker’s insightful quotes about business. This is one of my favourites: “What gets measured gets managed.”
Drucker’s quip encapsulates the data-driven mindset which is quite topical at the moment.The social data explosion and the growing awareness amongst marketers of big data and its huge commercial implications has made predictive analytics and data analysis a hot topic in both business and marketing circles. Who doesn’t want the ability to predict what customers want? (Tweet this)
Many organisations have already started their own internal analytics practices or have enhanced an existing one. Chances are the majority of these will have experienced moderate levels of success. But like most things in life, there are a few fundamental speed bumps along the way that can hinder your progress. I’d like to talk about three of the most important ones. If you can address these, then you stand a much better chance of translating your modest success into even greater results.
The main obstacle we see repeated time and again is lack of clear definition of desired outcomes. It’s a simple but extremely common problem. Better decisions from insight depend on how well the insight is incorporated within each process’s decision management framework. Now read that sentence again and ask yourself if you’ve properly defined your desired outcomes internally.
Another challenge is underestimating the need for a broader set of data. The proverbial “360-degree view of the customer” is in my view actually restricted to closer to 180 degrees. This is because it usually represents the known structured data that your company has collected over time. It’s usually not a 360 degree view because you typically lack external, unstructured, and often un-augmented data. In other words, you don’t know what you don’t know. Really gaining meaningful insight depends on how proficient your capabilities are for collecting, organising, and analysing data from varied sources – not just your own internal ones.
Historically, analytics teams have been set apart in their own department, because they’ve been doing something new or looking at things at a much deeper level, which is different to how many other parts of the organisation operate. This often leads to misalignment between the business and analytics teams (see obstacle number one!). For data to be applied effectively and with meaningful business benefit, it must be applied across internal departments and systems (Tweet this). And more importantly, those teams need to have an understanding of what each contributes.
Analytics can deliver powerful results, clear competitive advantage and translate into significant revenue. All the more reason then, to make sure you get it right.
What are your thoughts on this topic? Please leave any comments in the section below! And for more information take a look at this free report from the IDC.
Head of Marketing, SAP EMEA