Let's Stand Up to the Attack on Marketing-Mix Models

For Years, These Calculations Have Enhanced Marketing and Media Spend

By Published on .

Long-established marketing-mix models are coming under attack, with critics saying the approach was wrong all along and has only gotten worse with the emergence of digital and social media.

At the Advertising Research Foundation's Audience Measurement conference just concluded in New York -- aka "Geekapalooza," and I'm one of the geeks, so I can say that -- several sessions even proclaimed these models to be the enemy of worthwhile ad spending.

Really? Marketing-mix models are a statistical technique that explains how much of the weekly ups and downs of brand sales are due to various factors: media, promotions, pricing, competition, the economy, weather, etc. You then figure out the return on investment for each marketing activity. Some senior executives on the media side say the models systematically undervalue media and overvalue short-term promotions. To which I say: settle down Beavis, not true.

Barely a generation ago, the marketing and media ROI revolution, fueled by marketing-mix models, began a swift and steady march out of the basement of research experimentation. The models then made themselves at home on C-Suite couches and in boardroom briefing books.

To many, watching the near ubiquitous adoption of these models was akin to witnessing an Asian carp invasion. The waterways and ecosystems of marketing creativity became clogged and choked as numbers seemed to out-compete ideas. Sadly, in some timidly managed meetings, this probably really happened.

Yet the vast majority of these models (with their ROI figures and "sweet spots" on marketing-plan simulations) supported and enhanced marketing and media spend. I say this with conviction, having been a partner and CEO of a "mix shop" for over a decade.

Marketing-mix models helped show, for example, that marketers' infamous tendency to try to make their annual numbers by cutting ad budgets in the fourth quarter ("oh please, consumers won't notice a thing, sales won't suffer that much") was usually ultimately more expensive. In the long run, brands had to compensate in subsequent quarters to regain lost sales momentum.

Even the finance department, no fan of spending in general, would come on board. In one of many examples, the CFO of a top-10 retailer told me, in the presence of the CMO: "I don't hate marketing spend, I hate marketing waste. I'll march into the boardroom arm in arm with [the CMO] tomorrow and fight for more marketing money if I know it's going to grow the business."

The CEO of a multibillion-dollar global firm in the health-and-beauty sector proclaimed, "I want these insights for every major category in every key geography. When we add in our plans on new products, we'll have our platform for long-term growth. That's what I'm paid to do, not just 'make the quarter.' " Half a year later, armed with positive marketplace results, the CEO and the CFO jointly shared some of the findings with the financial analyst community. Not bad for a bunch of geeks and geek-ettes, eh?

I've seen first-hand how these models provide critical and compelling cases for media budgets, more than any other methodology I'm aware of. My firm (as well as my competitors) employed smart, sophisticated and highly principled practitioners with diverse professional and academic backgrounds. And if we were smart, our clients and their agencies were usually even smarter. They collectively subjected the results to exhaustive scrutiny and validation. Guess what? In almost every case the models were stable. They predicted what would happen in different spend allocations with amazing accuracy. Clients absorbed the advice, took action, reaped the rewards and came back for more.

This happened in consumer packaged goods, retail, pharma, financial services, telecom, what have you. Media and promotions were better balanced and worked with greater synergy.

Brands took advantage of free opportunities in media execution (spend level, timing, traditional vs. digital allocations), seasonality, halo effects (off-line affecting on-line sales and vice versa, one brand's media driving a sister brand sales) and portfolio management (how to lay down the money among all the brands an advertiser has for the greater good). During the great recession, mix models were front and center in answering the question of the day: "I have to cut spend, how I do it in the least damaging way?"

Mix models are all about improving the process. The models themselves also benefit from improvements in technique, methodology and smart application. They enhance judgment and strategy; they do not replace them. Such enhancements have and will help capture the longer term effects of media spend. Smart and successful practitioners have always expressed the media results as short- to medium-term calculations, and therefore inherently conservative.

Maybe my clients, my former competitors and I wouldn't find today's criticism so striking if it came from a new generation of marketing scientists. I'd tell myself and my peers "Hey, stand down, it's their turn." But this new Greek Chorus is comprised of folks that have (or should have) gray hair like me. They spend ten minutes bashing for every minute spent trying to revive their old alternative methodology. Moreover, none of us can recall these folks participating consistently and meaningfully in the rough-and-tumble ROI revolution we've been a part of for a quarter century. To now whine and imply that that all of us: clients, providers, agencies and consultants have been naïve and duped for decades is silly and, well, incredible.

Randy Stone is an independent consultant who works with media agencies, research companies and advertisers on marketing effectiveness and client/agency relations. He was previously CEO of the marketing analytics firm MMA (Marketing Management Analytics), then part of Aegis, for over 10 years.

Most Popular