Marketing Accountability: Do we have a plan for research and practice?
Earlier this week, the Wharton School of Business organized the first Informs Conference on ‘The Practice and Impact of Marketing Science’. The invitation-only concept was refreshing: let’s mix 50 ‘practical academics’ and 50 ‘thoughtful practitioners’ and see what comes out. I had lots of fun connecting with old friends and making new ones. Most impressive was Monday night: emptying all single malt whiskey in the local pub with four most excellent ‘practitioners’, while the conversation meandered from football and cricket to the relative benefits of MCI models and the best way to incorporate distribution in your marketing effect estimation! While the 1-on-1 conversations were, as usual, more memorable than most sessions, I fully enjoyed the practitioner-academic joint presentations, the Retailing workshop and the Practice Prize winners, who showed that working together on an important not-for-profit marketing issue can win over the audience and the jury without a ‘bells-and-whistles’ model.
Closest to my heart though was the workshop on Marketing Accountability, which I organized together with Neil Canter (Marketing Accountability Partnership). First up was Professor Russ Winer (NYU), who is the executive director of the Marketing Science Institute. He discussed interviews on marketing accountability with CFOs, one of which thought it was a joke to ask whether marketing should be capitalized rather than expensed: you can’t measure marketing’s long-term impact! In general, CFOs struggled to understand marketing, which they believed is focused on volume rather than cash flow or profits. Maybe a first step to explain marketing is to consider how the different levels of our metrics (customer mindset, product/market performance, financial performance and firm valuation) matter to different managers (marketing specialists, CMOs, CFOs and CEOs). The Marketing Science Institute received tens of submissions for its ‘Marketing Meets Wall Street’ initiative, in areas ranging from branding and CRM efforts to innovation and advertising. Moreover, the Boardroom project strives to develop ‘Marketing Asset Standards’ in metrics and methods. The $ 64 K questions are whether this can be done and whether it should be done, which generated a lively discussion.
Next, Neil Canter reflected on the poor state of traditional metrics: they focus only on the activation stage of the marketing funnel and do not link non-traditional marketing to ROI. Marketing mix analysis lacks long-term focus and has a tough time handling the lack of metrics and low penetration of new media. Continuous insight requires integration of data, next analysis, then dashboarding/business intelligence and finally advanced application. Currently, we are good at measuring the short-term impact of traditional media; the challenged is to look at the long-term impact of nontraditional media. In this respect, perfect is the enemy of the good: we need improvements in how we measure the effects of sponsorship, word of mouth, public relations, podcasts downloaded etc. What’s next is integration of metrics, behavioral metrics for the upper funnel and engagement. The currency for true cross-channel marketing accountability has yet to be determined. The solution will likely be a meld of multiple disciplines and require industry collaboration.
Suresh Divakar demonstrated how Avon juxtaposes the size of the marketing effect with the payback/ROI. Issues in the use of ROI to make decisions include modeling credibility and accuracy of incremental volume and the long-term effects of advertising, including those on brand equity. According to an MMA survey, only 23% of marketing executives and 6% of finance executives are satisfied with their ability to measure marketing ROI. Agreement is lacking on how to incorporate costs (gross or operating margin?), growth investment versus allocation decisions and how much information to give to Wall Street – especially for a privately held company. For instance, if Avon communicates marketing ROI to Wall Street analysts, they will expect so in the future and become alarmed with these numbers go down. Importantly, ROI measurement is just the first step, we also need to activate it into strategic plans and projects and ultimately make it part of the company’s cultural landscape. Critical challenges include ‘continuous’ activation, driving learning across multiple touchpoints in daily decisions and managing conflicting shortterm volume pressures versus long-term strategy implementation. Key learnings in Avon’s journey to activation include identifying agents of change, achieving full senior management support to reward good behavior, embedding the coefficients into the process, building guidelines and tracking into the Marketing Planning system, giving
road shows and on-the-floor Analytics engagement and placing people in local markets to lead implementation efforts. Finally, I discussed ongoing efforts to both bring finance to marketing and bring marketing to finance. The former requires us to tell senior executives when and where to increase and to cut marketing spending. Necessary steps include estimating the ROI and long-term effects of marketing actions, demonstrating how soon these effects will materialize and assessing the risk associated with the surge or cut. Based on these outputs, companies can design a system for scaling up certain marketing actions to fully exploit the short window of permanent benefits, and for scaling down or even killing campaigns that are no longer effective. However, bringing marketing to finance is at least as important, and has received far less attention from marketing practice and academics. We need to demonstrate that ‘soft’ metrics (e.g. awareness, brand trust) are leading indicators of company performance, preferably in marketing dashboard linked to financial consequences. In a recent paper, I compare three methods to narrow a set of 99 metric candidates to a manageable 10, based on the best out-of-sample prediction accuracy. Estimation of impact size and wear-in time reveals that different kinds of variables predict sales at distinct lead times, allowing management time to turn around the situation before performance itself declines. More controversial is ongoing research, together with finance professor Eric Ghysels, to demonstrate that Wall Street’s feedback on marketing policies is not always right, and thus should not be the only compass (or CEO compensation benchmark) to guide marketing decisions. We are the first to empirically demonstrate that, while poor stock performers engage in more risky product exploration, the stock market later does not reward, but rather punished such behavior. More generally, marketing managers and researchers, exposed to fashions and limited rationality of consumers, are naturally skeptical of the (strong version) of the efficient market hypothesis (EMH) that the firm’s current stock price is an unbiased estimate of future profits, and thus should not take Wall Street’s reaction always at face value. For updates, see my website: http://mba.tuck.dartmouth.edu/pages/faculty/koen.pauwels/