The Road Ahead for Marketing Measurement
Michael Wolfe
☆ CEO at Bottom-Line Analytics LLC | Advanced Marketing Analytics & Effectiveness Modeling
The Path Forward for Marketing Effectiveness Measurement
A few years back, I wrote an article entitled “the Death of Marketing-Mix Modeling as We Know It”.?This article warned of the demise of a tool and capability that had once become “the gold standard” of marketing measurement.?At the time, companies were starting to question its usefulness, the lack of significant innovation and its inability to adapt to the digital transformations going on in the commercial world.??As a result, ?new measurement tools such as Multi-Touch Attribution (aka, MTA) models came on the scene, which, were able to see how the digital flow of traffic affects sales conversions & could be precisely measured.?Through all of this, despite its digital-only focus, MTA became the measurement tool of choice for many companies
The whole marketing measurement proposition is once again at a crossroad.?The disappearance of digital cookies has pulled the rug out from MTA.??Do companies go back to the old-time marketing-mix model or is there a better path forward? ?Here is my vision of where this should go.
The Principles for Better Marketing Measurement
1.???Advertising measurement must inform both customer acquisition and customer retention marketing.
?a.???In other words, media measurement needs to focus on both short and long-term effects. Traditional marketing-mix modeling has been mostly focused on short-term effects of media, and hence, has often pointed management toward excessive short-term business results only.?The inclusion of long-term measurement puts more emphasis on brand building and customer-retention marketing.?In addition, including long-term ad effects has an outsized impact on the financial returns or ROI of media spend, which can lessen the propensity for the sometimes rash & unnecessary cutting of media budgets.
?2.????Advertising measurement needs to address the effectiveness of the advertising message or creative.??That is because 40-60% of the lift an ad produces can be directly linked to the message or creative quality of its ads.?Not all ads are equally effective; and marketing measurement needs to reflect that.
?a.???The effectiveness of advertising is a function of the quality of the message, its ability to persuade and how it engages its customers.?Many companies spend lots of time and money doing media copy testing and these tests can be built into predictive marketing models to enable a brand to truly monetize the value of their marketing communications.
??3.???Advertising audiences are exposed to multiple brand messages (e.g., multi-touch) across multiple media channels & measurement must account for this journey.?Advertising measurement must be designed to account for the multiple touches, interactions, and synergies across all media channels.
?a.???As a result of bias in regression & econometric methods, many believed traditional marketing-mix models to be flawed by not correctly accounting for media attribution covering multiple exposures to different media (multi-touch).?This led to the development of MTA or Multi-Touch Attribution models. While these models were solely focused on digital media, they became popular and very useful. Now, with the death of digital cookies, MTA is falling out of favor. ?Many companies want to know what is next.
?b.???Across time, new advanced analytic methods have come along, such as machine learning, Bayesian models and AI, because these can measure the interactions and synergies across different media. Some of the ?discoveries from these methods have been amazing, showing there is a large “synergy” operating between digital and broadcast media for many brands.?These new methods will likely usher in a new hybrid & type of marketing-mix model.?This is the direction marketing measurement is likely to take.
?4.???Advertising also interacts with other non-media business drivers and these interactions need to be measured and understood.?Marketing measurement must bring together how advertising is affected by and interacts with other business drivers.
?a.???For example, we have found that increasing advertising effectiveness often lowers a brand’s price elasticity; and this often happens when a brand is undergoing an improvement in its quality or when its advertising messaging is resonating more with customers.
b.???We have recently found that high levels of inflation cause some brands to increase their price elasticity; and this is often a bi-product of consumers switching down to lower-priced store brands
c.????Most of these interactions occur when there is a shift in the “customer brand experience (or CX).??This tells us that a CX measurement needs to be part of the larger marketing effectiveness model and equation.
领英推荐
?5.???To truly optimize marketing spend, effectiveness measures for media and other marketing must be made at the lowest level of detail.
?a.???Dynamic weighting can be adopted in predictive models which will enable valid effectiveness measurement of 500-600 separate media placements.?By doing this, marketers will have a greater opportunity to improve their bottom-line performance.
?6.???Marketing effectiveness measurement must be laser-focused for generating tangible business results and accelerating growth for the brands it serves
?a.???This is a marriage between the mathematical optimization of operations research and marketing.?By building simulators and monetizing all the inputs, companies can expect a growth dividend from 3 to 12%, without requiring an increase in their marketing budget.
7.???Marketing measurement must be timelier and push the envelope toward “always-on” and real-time results.
a.???This would be invaluable for saving money and enabling management to make more informed course corrections in business strategies & tactics.
?Which Direction is the Road Pointed?
?`I believe the direction that marketing effectiveness measurement is pointed is changing.?We can only hope that the destination is better than where we currently reside. This is the Manifesto and direction that I believe will make things better.
?1.????Marketing Measurement Models must account for short and long-term business drivers
2.????Marketing Measurement Models must account for and measure the qualitative aspects of marketing, including media creative effectiveness
3.????Marketing Measurement Models must connect all media channels, including digital and traditional broadcast; and connect with non-media drivers such as price and customer experience.
4.????Marketing Measurement Models must enable higher detail in its results to enable true optimization towards more effective budgeting of precious marketing dollars.
5.????Marketing Measurement Models must be better informed by data timeliness and move towards a more “always-on” and real-time exercise.
Michael Wolfe, [email protected]
Hi Michael - cool article. I think the point about price elasticity is underplayed in that for me, MMM needs to progress beyond marketing in the narrow sense. Consumers don't seperate price from advertising when it comes to learning about the proposition. MMM practitioners (should) know a lot from their analysis re pricing and promotional activities. These can have a much larger impact on the bottom line so I believe we need to be smarter in highlighting those impacts. Re the longer term value, I have seen a huge shift here driven by many trends you highlight. MMMs focusing on just acqustion may be OK but as you say they miss a much bigger picture - especially that which is driven by customer experience of the brand. For me timeliness, trust and understanding of where MMM fits into business processes are the key issues i see which practitioners need to work on. Speed can be solved in a number of ways. Trust stems in part from performance - often predictive performance. That takes time but is (in my experience) the key factor in helping organisations to embrace MMM. MMM's place in the business process differs of course but it clearly goes well beyond media planning these days.
Strategic Analytics and Insights
2 年Great article! ?? A dynamic, unified approach to marketing mix modeling is necessary. To win in the future you need to be able to understand how individual channels, media publishers, campaigns, creatives and non-media factors like brand, competitor media spend, inflation or even the war in Ukraine is impacting short and long term sales. You also need to be able to quantify the synergistic effects between all the different explanatory variables. Preferably this should be a single framework that is reliable, accurate and extremely actionable.
Consulting at DSource | Lecturer - Bilgi University
2 年Brilliant & thought provoking. Thanks for sharing Michael. A single framework to address all of these points would be the ideal solution ??
Business Analytics Lead | Corporate Finance | Machine Learning | Data-Driven Decisions | NLP | Applied Analytics & Statistics | Marketing Mix | Pricing Analytics | Forecasting | CPG/Restaurant | Test & Learn
2 年I do believe you still need to use both tools, which will help you to calibrate marketing contributions!
While Advertising is still there, in Digital the talk is also about Brand content which takes many forms : influencer stories/activations, brand dedicated programming, sponsorships... How do we get 'content' accounted for in MMM and experiential activations too?