Structuring Your Metrics for Cross-Campaign Comparisons
By Satnam Singh
Could the pressure be any more relentless on marketing departments to increase sales these days? I doubt it. CMO's want accountable marketing that provides insight beyond the usual calls, clicks, and sales. They increasingly demand that campaigns be comparable across the enterprise and supported by an analytics infrastructure that helps them make decisions that increase return on marketing investments.
Why is this so difficult?
For one thing, the measurement methodologies of different programs speak different languages. Branding speaks “awareness and perception” while direct response speaks “cost and profit.” Another is that individual program analyses cannot effectively account for interactions across channels.
What we need is a translator, an enterprise-wide, channel-agnostic structure that provides a single, comparative measurement. Such a system can help maximize ROI by identifying important interactions we've likely never noticed before. We must be able to answer such questions as, “Which campaigns create leads, which convert these leads, which do both?” and “What are the effects of campaign interaction on sales?”
So how does one create a system that delivers this? It's not as daunting as you may think. Here are the answers to common questions that arise when discussing this issue.
Q: Do we need to discard use of individual program metrics and analysis?
A: No. While enterprise-wide analysis is best, individual program analyses are still helpful.
Individual analysis is great for answering tactical questions, such as which performs best, the challenger or the champion? In these situations, a focus on the single program enables optimization of its different components.
However, metrics derived from such analyses do not provide comparability between programs. It's like trying to compare the value of different currencies. Sure, the value of 10 U.S. dollars may be less than 10 euros, but they may have a higher purchasing power because items are cheaper in the U.S. than in Europe.
A comparable metric accounts for these variations and is, therefore, preferable for making strategic decisions about such things as the allocation of marketing investment across the enterprise.
Q: Will the lack of a unified, standardized central database prevent the development of such a system?
A: No. Although the availability of standardized, centralized data is best, even a shoebox full of receipts can be used.
While a centralized database is ideal, in the hands of an expert statistician, anything that provides information on marketing investments combined with sales generated is enough to deliver valuable insights and development of the comparable metric. (Fortunately, none of our clients have presented us with a shoebox full of loose records -- knock on wood.)
We've found that clients don't often have centralized and standardized data on marketing programs and consumer response. But with the application of econometric modeling techniques, we have nevertheless been able to develop a comparable metric and increase ROI by double digits.
Q: Do we have to begin across all programs at once?
A: No, though it's a good idea.
Enterprise-wide analysis delivers the benefit of optimization of all programs across the enterprise. Most organizations have structures delineated by the channels used (direct mail, interactive, mass media, etc.) and the management complexity of different programs (direct response, brand awareness, etc.).
Analysis and development of a comparable metric for a subset, such as all direct mail programs, can deliver local optimization for the programs included. We've found a good compromise is a phased-expansion approach, which eventually leads to development of an enterprise-wide analysis. The only drawback is that it takes longer.
Q: What will the ability to measure everything in a comparable way do to the culture of my marketing department?
A: Expect increased confidence, transparency, and collaboration.
An enterprise-wide comparable metric gives you common ground for discussion and decision making, but more is required than just the application of advanced econometric techniques.
Culture shifts always take place when analysts take a seat at the marketing planning table. They begin by providing input into marketing briefs for program testing and measurement, and bring different departments together for discussing program interaction and profitability. In our experience, executive sponsorship is required to ensure collaboration between managers whose programs have significant interaction. Careful leadership is required to decide future investment questions about programs that are found to be inefficient.
It is certainly an exciting time for marketing and analytics given the emergence of new communications channels and the ever-increasing availability of data for analyzing consumer behavior.
Satnam Singh is Vice President of Marketing Analytics.
This article originally appeared in iMedia Connection, July 30, 2009.