Media Relations Measurement: The Next Step
3rd February 2007 by Eric Bergman, ABC, APR
This is the third in our series on media relations measurement and features an interview with Andrew Laing, president of Cormex Research Inc., a Canadian media content measurement and analysis firm.
Andrew believes there are three levels at which media relations measurement can be conducted: exposure; awareness; and attitudes.
Exposure is precisely what the name implies. This is similar to Media Relations Rating Points (MRPs), which we discussed in the last post on this series, but with an important difference. Andrew believes that exposure should not only measure where an organization’s name, products and messages appear in the media, but should include an analysis of demographics. In other words, were the publics important to the organization exposed to its messages?
“This involves understanding the demographics of the media in which your message is getting across,” he says. While exposure is relatively easy to obtain and relatively easy to measure by line counts, blog hits and messages conveyed on radio and television, Laing believes that exposure — in and of itself — is a weak goal.
His second level of measurement is awareness. He believes it’s important to determine how aware publics are of the messages that the organization is conveying through the media, which could include an awareness of the organization, the product, the issue or the messages the organization is attempting to convey.
As a step up from exposure, awareness brings a higher level of complexity. It is also more difficult to quantify because publics may be aware of messages through means other than what was presented through the organization’s news coverage. This means that measurement must be carefully done and must account for what Andrew refers to as a higher level of noise — other messages (such as advertising, word-of-mouth or other marketing programs) to which consumers or publics may have been exposed.
The third level of measurement is attitudinal measurement. This means determining whether recipients of the messages have been influenced by news coverage to have favorable or unfavorable attitudes toward the organization. To do this effectively, it’s important to survey attitudes before and after the measurement period.
Andrew believes that behavioral measurement is beyond the reach of most media relations activities. “In my view, it is very difficult for PR practitioners to attribute their (media relations) activity to behavioral change,” he says. “Behavioral change is a very complicated thing to determine. Media research, which has been conducted for decades around the idea of media effects, has largely abandoned the idea of trying to link media exposure with direct behavioral change. There is so much noise — so much interference — into what’s prompts a consumer make that choice.”
You can view a sample Cormex Research report conducted for the University of Alberta.
February 4th, 2007 at 7:06 am
There’s no doubt that Andrew is one of the smartest people in media measurement but let me suggest that there is a fourth element of media measurement beyond exposure; awareness; and attitudes: and that is “behavior.” And it need not be more involved than measuring the first three.
The method to which I refer is known as “marketing mix modeling,” or MMM, which is a statistical analysis that looks at the effects of advertising, price promotions, direct marketing, trade promotions and PR among any number of other factors versus sales within a particular place and time-frame. The results show the extent to which each marketing agent delivers sales — either in isolation or in combination.
The marketing mix modeling programs in which we’ve participated show that PR delivers the best return-on-investment (ROI)…as much as $12 for every dollar invested (vs. $1.20 for mass-marketing advertising and a -.25 on the dollar for price promotions). To not take the media measurement through to this outcome is to handicap the true power of PR.
What is most interesting about MMM is that it measures marketing effectiveness even if different forms of marketing are ocurring simultaneously; it shows the relative performance of each marketing agent so that investments can be shifted to what’s delivering the best ROI; and it eliminates people’s inability to remember the source of their awareness. Most people can’t remember what they watched on TV last night let along what they might have read, heard or seen a week ago…and that’s where the “sequential marketing process” (exposure leads to awareness, awareness leads to attitudes which lead to behavior) begins to break down. One has to be able to segment each stage of the sequence in order to know what’s working…if people can’t remember, you can’t document the sequence.
Marketing Mix Modeling eliminates the weak link — PEOPLE — and seeks to understand marketing and PR effectiveness by statistically linking marketing activity data with expenditure data with sales data.
The end game in all this, of course, is to uncover the impact of media relations on an outcome, the most of important of which are sales, stock price, etc. Sequential research and evaluation doesn’t accomplish this in the way that marketing mix modeling does and it needn’t be any more involved or expensive than other approaches.
February 5th, 2007 at 4:32 am
Mark,
Thank you for those insights. It’s my turn to add the disclaimer, however.
Please do not try marketing mix modeling at home. And please, do not apply a multiplier of “12″ to your advertising value equivalencies, based on what is written above!
February 5th, 2007 at 10:13 pm
Eric,
I think much of what Mark is talking about above is what you and I talked about in the first podcast in this series. However, I do not refer to these methods as “marketing mix modeling.” To me, the method used is simply the science of econometrics (or advanced statistics) applied to the business challenge of marketing measurement.
A major reason I don’t call the method “marketing mix modeling” is because I typically include environmental variables in the models I build for my clients. Environmental variables are those variables that the marketer cannot control — unemployment claim filings, temperatures, interest rates, exchange rates, changes in equity market indices, changes in commodity futures prices, etc. — and therefore are not part of a “mix”, per se, but reflect the reality of the environment in which the marketer is competing for consumers’ attention, time, and dollars (or whatever currency). I’ve built models where only the environmental variables had any statistically significant effect on sales — the marketers, regardless of how much money they spent or in what channel they spent it, had no significant effect on sales. It’s easy to increase Marketing ROI in that situation — stop spending money on marketing! At least until the environment changes to one more favorable and responsive to marketing.
In short, I believe that a marketing measurement practitioner should not simply apply a standardized “marketing mix” model to every client’s marketing situation. Instead, he or she should study the client’s unique business and include not only marketer-controlled variables but also relevant environmental variables (where feasible) to capture the effects of variables outside the control of the marketer that are also likely to influence consumer behavior (e.g., make a purchase). To leave these environmental variables out of a model of a client’s marketing situation is to mis-specify the econometric model and thus introduce error into the model before measurement and/or analysis even begins.
Finally, I agree with your disclaimer above - Mark’s quoted results are likely from a study of one client’s marketing initiatives and the results they achieved. Again, each client is different. Mark’s results likely apply to his client (or clients) only. A person should not read about Mark’s results on this webpage and then go and try to apply his results as some rule of thumb or heuristic in their own situation. Such a (mal)practice could lead to erroneous — and costly — management decision-making. Instead, take the time to build your own models, of whatever form, that “model” your unique situation best and then use the results of those models to make informed decisions about your unique situation.
February 6th, 2007 at 4:01 am
Logan,
Very insightful, as always. Thank you for your input.
February 8th, 2007 at 11:06 am
Logan and Eric,
Thanks for providing a platform for this dialogue. Delahaye also factors environmental factors, to be sure, including the weather which, for example, might accelerate the sale of beer during hot summer months or rain, which might accelerate the sale of movie tickets (both of which we’ve found to be true in models of our own). Whether this form of statistical analysis is known by “marketing mix modeling” or econometrics, the bottom-line truth is that the “science of PR” is exploding, that these analyses will dictate funding and programming and those who fail to recognize the implications are at great risk.
And since everyone has offered a disclaimer on my post EXEPT me, let me say that the results contained in my post are for those programs which were processed through a rigorous series of anayses. They can not be projected at will to all marketing and PR programs. However, the consistency with which PR delivers the best ROI after dozens of cases in a dozen industry categories should make us all more comfortable when claiming “PR works.”
February 11th, 2007 at 5:51 am
Mark,
Thoughtful, as always. I didn’t mean to overstep my bounds with the disclaimer. I was just actually thinking of our conversation, which is being posted this week.
Cheers,
//eric