Allocating marketing budget using marketing attribution and marketing mix modeling (Part 2 of 2)
In the previous blog, we discussed various aspects of multi-touch attribution and how it is more relevant for heavily online businesses. In this blog, we will consider a different methodology applicable to companies that employs both online and offline marketing channels – Marketing Mix Modeling (MMM).
MMM methodology originated few decades back when marketing was not so trackable. It has been used especially by consumer packaged goods (CPG) industry. Such businesses used media advertising and heavy couponing. MMM involves performing econometric analysis on several years of data and establishing relationship with a dependent variable (sales). The resulting equation can be used for scenario analysis with different marketing investments in different channels and how they affect potential sales.
Need help with Marketing Attribution and Marketing Mix Modeling? Want to leverage the latest in data science and marketing operations to turbocharge your marketing performance? Contact us for a FREE consultation!
Data needs of MMM is extensive, involving in two years or more of past data:
- Advertising quantity and amounts – by time and geography
- Promotions of all kinds (coupons, rewards, etc.)
- Macroeconomic factors (unemployment, economic growth)
- Introduction of new products
- And more…
The econometric equation derived (that predicts sales) is a bit of a black box. However, over the years, MMM has been widely accepted in traditional marketing departments as a dependable methodology.
MMM provides insights on which marketing channels are the most effective (elasticity) and how yield curves indicate diminishing returns by channel for higher investments. MMM also captures the cumulative effect of channels, one reinforcing the benefits of other (“1+1 is more than 2”).
MMM model need to be refreshed every two years or so to keep up with dynamics of marketplace.
Comparing Multi-Touch Attribution (MTA) and MMM
One major difference between the two methodologies is this: MTA looks backward from conversions and see what marketing channels and touches contributed to it. MMM on the other hand, works forwards from the marketing investments and see how they lead to sales.
As discussed, MTA is ideal for online businesses that are trackable. MMM is ideal for the offline (non-trackable) businesses.
However, in many cases, businesses use both online and offline marketing channels. eBay for example, spends heavily in Google advertising, but also on TV advertising and even on radio. In situations like this, it would be worthwhile to try both methodologies. Analysis that look forwards (MMM) and backwards (MTA) together will provide extra insights that can be very beneficial.
There are new methodologies being developed that connects MMM and MTA, but currently nothing has emerged as widely-accepted. For the time being, marketers should use their best judgement in determining the real marketing channel attribution and therefore, budget allocations.