Periodically, I feel the urge to rant about the current obsession with marketing ROI. I say obsession, not because marketing ROI isn't a very important measure of marketing performance, but because marketers now seem to feel compelled to calculate the ROI (or a projected ROI) of almost every marketing activity - even when the ability to accurately measure ROI is questionable at best.
I published my last rant on this topic in January of last year, and Stop Trying to Measure Marketing ROI has become one of the most popular posts at this blog. I won't rehash all of the arguments here, but the biggest challenge in measuring the ROI of an individual marketing program is revenue attribution. In order to calculate the ROI of a marketing program, you must know how much incremental revenue the program produced. If you can't accurately attribute revenue to a marketing program, you can't calculate an accurate ROI.
I was thinking about this topic when I came across a post at the Harvard Business Review Blog by Alnoor Ebrahim, an associate professor in the Social Enterprise Institute at the Harvard Business School. Ebrahim's post discusses how three "social action" organizations measure the performance of their programs. The focus of the post is whether the organizations only measure the immediate outputs of their programs, or also attempt to measure ultimate impacts or outcomes.
For example, Acumen Fund is a venture philanthropy fund that invests in social enterprises in Africa and Asia. Its primary social metric is the number of lives reached in poor markets. If Acumen invests in a company that manufactures anti-malarial bed nets, it will count the number of nets made and distributed. Acumen does not try to measure ultimate outcomes such as reduction in malaria or improvements in health, because it believes that measuring ultimate outcomes is too complicated, expensive, and impractical.
Robin Hood Foundation is a grant-making foundation whose objective is to fight poverty in New York City. When Robin Hood makes educational grants, it first identifies a set of results that can be easily measured - increased school attendance, scores on standardized tests, and high school graduation rates. Then it attempts to find third-party research studies that correlate these near-term results to expected lifetime earnings or quality of life (the ultimate desired outcomes). Robin Hood uses these studies to estimate the ultimate benefits of the programs until direct measurement (or better research) is available.
Professor Ebrahim argues that organizations must be realistic about measuring ultimate impacts: "Surely measuring impact matters but we need to be realistic about the constraints. It requires a level of research expertise, commitment to longitudinal study, and allocation of resources that are typically beyond the capabilities of implementing organizations. It is critical to identify when it makes sense to measure impacts and when it might be best to stick with outputs - especially when an organization's control over results is limited and causality remains poorly understood."
So, what does this have to do with marketing? I would suggest that the measurement challenges facing marketers are similar to those faced by these philanthropic organizations. Marketers would like to quantify the impact of every marketing program on revenue growth (the ultimate desired outcome), but that may not be realistic in some situations
In today's B2B marketing environment, prospective customers will be exposed to numerous marketing messages and programs over the course of their purchase journey. On top of that, for B2B companies that offer complex products or services, personal selling plays a significant role in driving new sales.
The issue is: How do you accurately attribute revenues across all of the marketing and sales activities that play some role in the generation of those revenues? With the use of extensive, longitudinal testing and marketing mix modeling, it may be possible for a company to arrive at a reasonably accurate attribution of revenues. However, these techniques require significant expertise and can be very expensive to use. As a result, few companies go this far. Research by the Lenskold Group indicates that only 11% of companies use test and control groups, and only 3% use market mix modeling. Without these techniques, it can be all but impossible to accurately attribute revenues in a comlex demand generation environment.
For most companies, the more practical approach is to measure the outputs of individual marketing activities and to correlate those outputs to revenues without trying to attribute a specific dollar amount of revenue to each activity. With this approach, you can judge the value of an individual marketing activity without needing to use arbitrary revenue attribution to calculate ROI.
For a thorough and less "ranty" discussion of this topic, I recommend that you take a look at this recent post by Jon Miller at the Marketo B2B Marketing and Sales Blog.