Predictive Analytics is for Experts only! – Really?
You don’t need to be a scientist to boost your business with applied mathematics
On 22/9/09 SPSS Inc. announced a new certification process to confirm an individual’s expertise with some of their statistical solutions. “Look at this”, I thought “sophisticated software still requires experts to unfold the value they can provide”. Being a physicist by background, I like it how applied mathematics can improve business. However, not everyone sees beauty in algorithms or is interested in statistics.
It’s true – predictive analytics requires deep understanding to be business relevant. Traditional analytics is based on historical data. Although the interpretation of such data can be very tricky as well, there is nothing wrong with the data itself – it’s fact (let’s assume that basic data management is established). Predictive data however, can be totally wrong if the applied model doesn’t fit to the business question it is supposed to answer. Experts are needed to make sure that the predictions are more relevant than a look into the crystal ball. But where do these experts need to be? Do all companies using ERP software have development experts for the solutions they are using? Certainly not, and in general users are advised to keep the system ‘clean’ from modifications and apply enhancements with care. Customers trust ERP vendors that their products are compliant with the rules and the same applies to analytics vendors as the market matures and ‘advanced’ analytics are becoming ‘standard’ analytics.
There are three different users for predictive analytics
• Analytic experts understand how predictive models work and can ‘enhance’ them for specific business processes if needed. Larger organizations will want to have such experts in-house but vendors (and service companies) will offer pre-configured forecasting and decision support solutions for different verticals and business processes that will eliminate the need for on-site experts.
• Analytic users are actively working with the analytic tools but wouldn’t care much about the complexity of the underlying models. These users understand and interpret the results and use them for their business decisions in their daily work. In order to increase adoption of analytical tools across the business, usability for these users is key for success.
• Analytic consumers are making use of the output from advanced analytic tools, often without actually knowing where the information is coming from. This could include a call-center agent who gets an alert on-screen while talking to a customer or an ATM machine that reacts to forecast information while processing the request from a customer.
Experts are needed to correctly apply predictive analytics, but everyone can use it to boost business efficiency!
Please leave a comment or contact me directly.