Traditional models of PR measurement no longer work (i.e. AVEs). Digital has become intrinsic to the PR and communications landscape, and diversification is forcing measurement to become more innovative than ever. Both PR agencies and in-house communications professionals are struggling with this shift.
“As planning lead for public relations, I constantly deal with the need for better data to define a business problem and set ROI (Return On Investment) measurement beyond free PR coverage,” says Marion McDonald managing director, strategy and measurement, Asia Pacific Ogilvy Public Relations.
For the past several years industry organizations, PR professionals, and experts have attempted to experiment with new models for accurate measurement of PR outcomes and the value of corporate communications to an organization. On a broader level, this presents challenges because traditional media relations has become a singular (and increasingly narrow) part of the overall scope of public relations. As the industry continues to broaden and diversify, areas that were once considered only creative advertising tend to overlap with PR and communications. In a time where measurement is increasingly critical, it’s becoming more and more complex to do.
A new approach
Have you heard of Media Mix Modeling? Sometimes called MMM, it’s an analysis technique that allows marketers to measure the impact of their marketing and advertising campaigns to determine how various elements contribute to their overall goals. Media mix modeling provides insights that allow marketers to refine their campaigns based on several factors. For example, marketers can create campaigns that will better drive engagement and sales based on things like consumer trends and external influences.
How does MMM work?
MMM leverages aggregate data. As part of this function, MMM is able to evaluate a wider range of channels, whether traditional or digital. It also allows marketers to factor in external influences such as promotions, seasonality, or other variables that were previously not addressed.
MMM draws “market-influencing” data from disparate sources, applies advanced statistical analysis, and provides insights into the efficiency and effectiveness of ongoing marketing programs. Leveraging traditional variables like sales, advertising Gross Rating Points (GRP), or a content analysis of news coverage, researchers are able to examine data to create a more complete view of the marketplace and draw a conclusion about the impact on sales.
MMM seeks to quantify the impact of individual marketing activities on sales volume. It does so by accounting for the effects of controllable elements like advertising and pricing, as well as semi-controllable things like PR or sponsorships. Layered with those variables are uncontrollable external circumstances, and together it’s possible to calculate the relative and absolute returns of each particular marketing activity. This enables marketers to determine their best budget allocations among different marketing platforms, because MMM takes each communication expense into account. This model helps marketing leadership to understand at a very real level what parts of marketing spend are being wasted.
The downside? The model is only as good as the inputs it receives. Garbage in, garbage out, as they say. And it can be very difficult to take into account every possible factor that could drive sales. That’s because PR professionals are not trained in statistical modeling, and people with that particular mathematical skill set are not very familiar with PR in the marketing communications mix.
So what does Media Mix Modeling mean for your organization?
Media Mix Modeling could provide some of the most accurate and robust insight into marketing impact. However, a number of factors need to be met to truly determine ROI, allocate future spend, and create sales forecasts. To learn more about how MMM works and if it might be worth exploring, watch for our second blog on the topic. In our next post, we’ll share examples of how MMM works, how it’s different from just data-driven attribution, and what is needed to implement MMM.