Marketing Attribution & ROI Measurement Needs become an Affordable Appliance

This month Jim Nail at Forrester wrote a report called "Marketing Measurement Metamorphoses", which nicely summarizes how marketing mix modeling and attribution are converging. I was interviewed for the piece, and our SaaS based attribution solution SpendMetrix® was profiled briefly as well. 

We believe that many marketers who are doing a mix of marketing and advertising across digital, CRM, traditional and other media have basically been shut out of the opportunity to do thorough attribution and ROI measurement that spans all media. This is in part due to the high cost of custom modeling efforts, the difficulty of corralling the fragmented data together, and the organizational forces which resist measurement in marketing. 

When a new solution has a big ticket price it can get bogged down with procurement and IT involvement. Add to that the professional risk that is often perceived in advocating for measurement within the marketing group, plus the reluctance of ad agencies and you can see why there is inertia around marketing attribution and measurement. 

Our SaaS solution, SpendMetrix®, grew out of 15 yrs of building bespoke marketing mix models for small to mid sized marketers.  Almost all of them are very well known brand names, but many don't have large advertising and marketing budgets. It's not unusual for a multi billion dollar client to spend less then $20 million / yr. on advertising. In South Asia, where marketing budgets are often run country by country this is very common. And in every case we worked on, the gains to marketing spend efficiency from our work were 15 - 25%, with as much as 3 - 5% in top line lift in sales. 

We designed our SaaS solution to overcome the hurdles and get ROI measurement and optimization through an automated solution. The core of the solution is our TreeOpt® algorithm which automates both prediction (attribution) and ROI optimization in a very flexible and efficient way. It captures deep multi way interactions or media synergies, non-linearity’s, diminishing returns to spend, and offers optimization within constraints.

SpendMetrix® also provides all the same marketing mix and attribution metrics that come out of these technologies, including baseline vs. incremental sales lift, incremental ROI's, and granular reallocation advice. And because it integrates, web traffic, survey, marketing, and lead and sales data, SpendMetrix® can also be "pointed at" non-sales metrics like brand awareness, and web site traffic and engagement metrics as well.

We have found in tests against past client data, and against simulated data where we know the true ROI and response coefficients, that we can routinely capture 80 - 90% of the gains that come from optimization using this out of the box automated approach.

 It turns out that most of the gains from measuring marketing ROI and reallocating marketing spend come from rank ordering spend options and differentiating the best (say top 10) from the worst spend options. These spend options can be as simple as City x Media Type x Creative but also drill down to very granular spend segments limited only by the granularity of the client data.

This is oversimplifying a bit, and our solution certainly does more than this, but at the end of the day optimization is about finding the top 10 -15 spend options and “doubling down” on them. This can be automated for small to mid sized advertisers, and arguably for larger advertisers as well.  By automating the modeling process, and by making it easy to corral the scattered, messy marketing data that exists today, we have made something practical and cost effective for marketing attribution, ROI measurement and optimization.

We have created a marketing attribution "appliance" and we believe that that marketing measurement world will continue to evolve quickly in this direction. The cost of SpendMetrix® depends to a degree on scope but in all cases it can be done for less than 0.5% of the client's marketing and advertising budget - in contrast to the minimum 6 figure, 1 - 5% of budget costs that are frequently associated with mix modeling and measurement solutions today.

If you would like to learn more please see the Forrester report, visit our web site www.mdscience.com or drop me an email directly. 

John Cripps, Ph.D.  - CEO Marketing Decision Science.