Unfortunately, PR has fallen into the trap of measuring what is easy to measure, even if it doesn't really mean much. Measuring impressions and ad equivalency prevail (see Erica Iacono's August 11 PRWeek column for an overview of the dangers of impression numbers), while what is most important to measure – the effects of PR on sales, revenue, and stock valuation – is rare.
The difficulty in moving PR measurement forward is the persistent myth that measuring ROI is impossible. It's not impossible, but it isn't easy. The measurement requires sophisticated statistical modeling to tie PR activity to financial value. Though not easy, it is very valuable.
Statistical models offer a fantastic way for PR pros to demonstrate the financial value of their efforts. When applied, statistical models consistently support what PR pros have always known, but have never been able to unambiguously show to C-level executives: dollar for dollar, PR expenditures generate greater long-term revenue than advertising and have a much higher ROI.
While the benefits of statistical modeling seem obvious at first glance, our clients often show trepidation when we suggest using them. In the past, only academics and some very large research firms have used statistical techniques to demonstrate PR's ROI. This has entrenched the belief that the analyses and results are too complex for business settings. But the reality is that the different models used to demonstrate PR's real value, like “marketing mix modeling” and “multivariate analysis” all produce an intuitive metric: the percentage of sales and revenue accounted for by PR (and any other marketing) activity.
The difficulties inherent in measuring the real value of PR through statistical modeling are numerous, but hardly insurmountable:
• The first challenge is the amount of different data types required to build valid statistical models. Data sources are more heterogeneous than has traditionally been the case in PR measurement. If you want to calculate PR's real value, you must look at PR and marketing expenditures, sales, and revenue addition to the traditional media metrics.
• You need to find the right partner to help you build the model. It doesn't need to be a small army of Ph. D-level statisticians (and will likely be a lot cheaper if you don't go that route). However, you will need to work with individuals who possess the ability to create statistical models and translate the results into language that both PR pros and the C-suite can understand.
• If you're in the services side of communications measurement, tell your clients to be patient. Collecting all the data for a good statistical model takes time. You'll likely need one or two years' worth of data.
Moving PR measurement beyond impressions will not be an easy task. Demonstrating PR's contribution to company value requires organizational commitment in terms of data collection and time, but the benefits are sizable.
Seth Duncan is a research manager and Nils Mork-Ulnes is a VP at Context Analytics.