There are many lessons to be learned from the Brexit referendum in the UK and the recent President election in the US. Since we are a marketing company, we will leave the political discussion to others and focus on an area closer to us – the analytics and forecasting.
These two events puts the spotlight on the difficulties with forecasting. Despite having the smartest people in the business, using the best methods and tools with unlimited budgets, they still failed. In both cases the final result was a huge surprise for us all. Not even the most advanced statistical forecasts, or combination of forecast methods, were sufficient to correctly predict the outcome. And it was not even a complex forecasting situation as such, there were simple choices – either Brexit or Remain, either Trump or Clinton. Even the dates were fixed, still hardly anyone managed to make a correct prediction.
If we translate this to a marketing scenario, we have some lessons to learn. The analytical situation in marketing is by far more complex, with more different choices, more competition and no time constraints for the market to make decisions. The budgets for marketing analytics are limited, as are the resources.
Despite the failures of forecasting the elections and despite the added complexity in marketing analytics, many marketeers are still dreaming about trustworthy predictions. Despite that we see many historic failures in analytics driven companies. Nokia is an interesting example, but there are many others. Companies that spend loads of resources on business analytics – and still fails.
We need to ask ourselves if business analytics, predictive analytics or whatever we want to call it, is relevant and worthwhile? If all the resources spent in political forecasting are not providing a correct output, why should business analytics be any better?
The response should probably be pragmatic. We should understand the limits of analytics, to learn how much it can be trusted. Don’t get us wrong, it is still valuable and should be a standard part in the marketing toolbox, simply because it is the best that we can do. It is better than doing nothing, advanced statistical analytics and forecasting is a learning process. It will get better over time and with more data, but still the results are often far away from being the truth. That is an important lesson to learn from Brexit and Trump. A big difference however, is that marketing predictions are rarely focused on such binary questions but rather used to indicate ways of improving and optimizing.
Another lesson to learn is to focus on keeping things simple and quick. Measure what you know is important, get the feedback quick and follow up on trends. In B2B marketing for example, many marketing teams are measured on the number of leads that are generated. In B2C marketing there are other basic KPI’s – focus on them. Don’t make it into rocket science and be sure to follow up daily or weekly.