Correlation and Causality. Words not often used by the communications industry. Initially because we didn’t need to. Now because we generally don’t want to.
It’s no secret that the communications industry continues to come under increasing pressure to demonstrate that communications efforts are contributing to the organization’s bottom line. This is where practitioners generally roll out a number of counter arguments not to measure.
The first is that measuring PR is a bit like catching water with a fork or counting a bucket of eels. I prefer to think of it as challenging but not impossible.
Counter argument number two, among practitioners, is that we can’t prove it was exclusively PR that did what we want (in a MarComm context, say) so there’s not point in trying. That argument only worked for so long among the consumer packaged goods companies that have figured this out—cracking the causality nut–with something called market mix modeling.
With a few exceptions, measurement in/for/of PR is really simply about adopting mass communications, sociological, and/or market research methods for our unique needs. The marketing and advertising side of the industry has been doing this for coming up on 100 years.
Consumer packaged goods companies, have, for decades, been tracking every sort of communication (I mean this broadly here: advertising, point of purchase, coupons, direct etc. etc. etc.) and PR, for a very tense while, was the only hold out. Well they started to track PR inputs and outputs as part of their market mix models, and two remarkable things happened:
- they were able to statistically isolate for and ‘prove’ PR’s unique contribution to the marketing mix, and (more importantly)
- were able to quantifiably validate a couple of age old assumptions: PR works and it often significantly outperforms other channels
Counter argument number three looks a little like this. “OK, so it can be done, but it must cost a fortune.” No argument there. It’s not cheap. But what’s curious is that if we can’t afford Market Mix Modeling (causality), then are we at least demonstrating a correlation (between, for example, quality and quantity of media coverage and public awareness)? Very rarely. Particularly in Canada.
Correlation NOT causality. PPoxy NOT PRoof.
With correlation, we’re saying that there is some statistically valid relationship between, say, coverage and awareness. (To be clear, though, we’re not indicating that one was exclusively driving the other.) How do we do it? Well, where market mix modeling involves sophisticated regression analysis and statistical modeling, correlation is a fairly simple—and single–statistical calculation. It’s called Pearson’s Product Moment. Wikipedia will have much to say about it, I’m sure.
Alan Chumley, Director of Communications Research, Leger Marketing, is an instructor of communications research in the PR programs at Ryerson and McMaster Universities, an associate member of the CPRS measurement committee, as well as an industry speaker, conference chair, and blogger: http://alanchumley.wordpress.com[ad]