I’ve been teaching a marketing metrics course recently. It’s a great opportunity to reflect on concepts that I often take for granted, topics that I am familiar with but don’t think about much in my day to day. For me, one of them has been statistical analysis.
I took several statistics courses back at university as part of my degree. I enjoyed the subject at the time, but I hadn’t actively used statistics in years. The exception is, of course, election time (I have a weakness for pre-election poll results).
I’d certainly not applied statistics as part of my marketing responsibilities, at least in a deliberate way. The companies I have worked for didn’t manage volumes of data substantial enough to require statistical analysis and large-scale modelling. Metrics were mostly compiled and assessed manually, and that was more than sufficient, thank you very much.
But today, as part of the innovation path led by Cloud computing, big data and the Internet of Things, things are looking very differently. The amount of data generated every day by all kinds of organisations would have been unimaginable a few years back. More and more data collection points are giving us vast amounts of data. The trick is that the only way to turn them into valuable information is to carry out some statistical analysis.
A/B testing has become commonplace, but to understand it it’s essential to have a basic understanding of statistical significance. You’ll need it to figure out whether any differences between your control and test groups are relevant or not. We aren’t necessarily after 90% plus confidence, 80% plus is probably enough, but you need the numbers to tell you what works and what doesn’t.
Marketing Mix Modelling (MMM) also uses statistics (more specifically regression) to identify the relevance of your campaign elements and determine which ones have performed best. Because there is an increased emphasis on marketing metrics, performance analysis and ROI, MMM is gaining relevance. MMM can provide many insights into the work of marketers, but it used to be expensive. These days, technological advances are eroding the high cost of drilling large amounts of data. As a result, MMM will become much more widely used.
If you work in FMCG, you’ll have been using them for aeons, and you’ll be wondering what the fuss is about now. But for anyone else working in marketing, using statistics as part of your marketing analysis will soon be a given. I am confident we’ll all be using statistical data on our day to day work, from detecting trends to forecasting to analysing campaign results, much sooner than we think. Adding a competent understanding of statistics to your skill set may sound like a smart move now, but it will be an essential requirement very soon.
Back to the classroom, I have to say that getting reacquainted with statistics has been a very pleasant surprise. Statistical analysis has come a long way since the clunky SPSS packages of my uni days. The Internet is making it incredibly easy to add statistics to any marketing campaign. There are numerous tools and calculators covering from confidence intervals and sample sizes to everything else in between. Even better are the many how-to videos, articles and tutorials on basic statistics. Go and check them out and start using them today.
Your marketing will be all the better for it.
[Photo credit: Lendingmemo]
