A Data Science Central Community
These two methodologies are often pitted against each other as competitors for the same Marketing dollars in today’s marketplace. In truth, that’s rarely appropriate.
On the surface both methodologies offer an ROI and an attribution for digital marketing vehicles such as online display, paid search, e-mail, etc. and thus create conflict by providing distinct answers. But once you learn more about them it’s like comparing an X-ray machine with a Microscope. They both provide a “view of the subject” and those views will “look different”, but it’s pretty clear which one you need in different situations.
Marketing Mix: Strengths and Weaknesses
MMM – What is it?
Marketing Mix Models are based on time series analysis. They attempt to measure all impacts on sales and put those influences in a neutral measurement system. The basic idea is to track the timing and changes in influencing factors such as advertising, pricing, and distribution and then correlate those with the corresponding changes in sales.
MMM – Strengths
Any data source that can be tracked over time can be incorporated into the analysis. This means that you can get an even handed valuation of all influences ranging from seasonality to marketing, short-term sales to brand building effects, product introductions to regulatory changes, etc. It’s like a full body X-ray. Presuming it’s done well, it’s accurate at assessing the proportions and magnitudes of the factors influencing sales. And like X-rays it provides an indispensable big picture view, without which you’re left with “exploratory surgery” once things have gone seriously awry.
MMM – Weaknesses
Modeling at the aggregate is great for big picture decisions, but like all big picture views the details can get a bit fuzzy as you drill down which limits that objective. It’s true that a good analyst can focus the analysis, use Bayesian priors, look at multiple data sources, etc. and like the doctor who has you drink a special dye that shows up on the X-ray, it will probably be surprising just how much can be determined from the analysis. But at the end of day, no matter the skill, every tool has its limits and MMM is an aggregate level analysis.
Digital Attribution: Strengths and Weaknesses
Digital Attribution – What is it?
The digital revolution has brought with it the ability to see which digital advertisements a consumer saw, clicked, and the sequence that preceded specific actions and purchases. Digital attribution is the process of using those click streams to predict online sales and attribute the portion of online sales caused by the specific marketing vehicles to the purchase.
Digital Attribution – Strengths
No other analysis is able to provide the detailed view offered by Digital Attribution. Alternative online displays shown to similar consumer groups can be evaluated, specific advertising venues can be compared, in short, the timing, the sequence, the content, and the location of online marketing advertisements, relative to specific audiences, can all be assessed and improved. It’s the microscope of digital marketing.
Digital Attribution –Weaknesses
Digital Attribution offers an unprecedented view into a previously unknown world, but like the microscope, the width of its view is very limited. Most cookies used to track online behavior are erased the same day they’re created, and many users do not accept 3rd party tracking cookies at all. Who can be tracked, and the length of online effects that can be measured, are both severely truncated. Perhaps more concerning are offline effects that are not tracked. Influences such as offline marketing, seasonal impacts, economic changes, etc. are normally omitted, and to the extent that they are correlated with online spending, these missing factors cause an over attribution to the online factors. Not controlling for important drivers always builds in significant biases, so while Digital Attribution is an excellent tool for many online comparisons, it is not as ideal for big picture views or comparisons with offline marketing.
Optimal Marketing Investments
Clearly both MMM and Digital Attribution have important contributions to make in optimizing the marketing budget and improving marketing ROI. Today the race is on to make the most out of what each methodology offers with the leading measurement companies vying to do the best job of offering clients an integrated and complete view that incorporates the strengths of each analysis.
Originally posted here.
David Young, Sr. Group Director, Neustar / MarketShare lives in Vienna, VA, and has worked in Marketing Analytics 20+ years.
We explored the same topic here earlier in Anayticbridge -
Thanks for the heads up.
Thanks David, I found this to be a useful succinct summary.
May I ask, would you consider Econometrics to fall under 'MMM' ? If not, can you explain where it would sit in relation and why ?