Picture the scene, you are sitting in a meeting room discussing your recent marketing campaign and asking yourself “What is driving my sales (or other KPI)?”
What are your options?
You could guess. It’s a time-honoured approach and, let’s face it, in the absence of anything else, it’s about all you can do. But, to do more than just guess, you could turn to Econometrics (aka “Marketing Mix Modelling”).
Econometrics, an advanced statistical technique stolen from the hallowed halls of academia, is ace at the all-important task of working out how many sales are being driven by each bit of media you’ve used (aka “marketing attribution”). And once you’ve got that (thanks, econometrics!), it’s a short step to working out your media ROIs and optimising your media plans to give you even more goodies the next time round.
That last step is particularly important. It’s all well and good knowing how your marketing performed in the past, but where econometrics really becomes valuable is when we use it to help shape the structure of future campaigns – which products you will support, with which media, and by how much – and forecast the likely outcomes based on something more than guesswork.
What does Econometric Modelling do?
In real life, everything happens at once. Some of it affects your sales (or whatever KPI it is you’re trying to move); some of it, less so. It doesn’t take much to see that in order to drive as many sales as possible, you need to know which things (“drivers,” we call them) are doing the work.
Enter econometrics. As much as data science and attribution has advanced in recent years, econometrics remains the only technique capable of measuring the impact of simultaneous drivers. By looking back over enough historical data for your media spend and other activity, an econometric model measures the correlations between what you did and what happened to your sales as a result. From there it calculates the effect that each activity in turn had on sales.
But what about seasonality, price, the weather, etc? They affect sales, don’t they? More than likely! And we can use the same econometric model to measure the effect of the non-media drivers on sales. In fact, we have to, in order to avoid over-attributing sales to the media drivers (beware the econometric model with only media in it). Handily, that yields sundry insights about non-media drivers that are often of interest and value to marketers.
We’ve discussed the what and why, now let’s explore the how
Econometrics takes a bunch of data (“variables”) and work out the relationships between them, by analysing their correlation over time. It doesn’t rely on tracking, claimed attribution, Google Analytics, cookies or any of that stuff. Instead, Econometrics combines elements of mathematics, statistics and economics – yes, really all of that. Don’t be daunted, though. Our forebearers have already figured out the theorems, solved the equations and, not least of all, made computers to help us out.
Did you know that Excel is building an Econometric model (albeit, a very basic one) when you use the trendline feature in a chart? However, as much as we all love Excel, you soon discover its practical limitations when it comes to building Econometric models in anger.
So, we use fancy computer packages, such as Eviews, R, SAS, Stata and so on. Believe us when we say we are very grateful for them. We’re old enough to remember the desperate hours spent in university exam halls, doing econometrics on nothing more than ProPlus and a pocket calculator.
It’s easy to get excited about a shiny stats package, but it’s not the main thing here. The people who build the models are known as “Econometricians” they are (or should be!) experienced analysts who know stats, love data and understand your business. With much application of elbow grease and long hours staring at screens, they test seemingly endless permutations of drivers to arrive, finally, at a model that both passes the statistical tests and makes sense. When you buy an econometric model, you’re also buying another 99 models that were considered, discarded or refined en route to the finished masterpiece that sits in front of you now.
This is just an introduction to give you a flavour of econometrics. There is much, much more. And a lot of it we spell out in our fabulous ebook, ‘A GUIDE TO Econometric Modelling for Modern Marketers.’ Download it for FREE to find out more.
Or perhaps you’ve heard enough and are thinking that this Econometrics thing might be just what you need to power-up your marketing efforts. Then, please get in touch