There are different ways to measure return on marketing investment but Media.Monks’ Arjun Ghosh says that only when multiple methodologies, technologies and skill sets are connected into one unified framework can marketing challenges be effectively addressed.
We have always seen recessionary fears spurring measurement buzz. In uncertain times, marketers turn to measurement when there is an urgency to justify the incremental value of marketing to the CFO, as well as to trim the increasingly long and complex tail of marketing activities.
We rarely go to doctors or dentists when we have good health and shiny teeth. Marketing measurement professionals are in the same category, with skills that should be used on a regular basis to anticipate and prevent problems but are sometimes only called upon in emergencies.
However, since many marketers are initiating or expanding their measurement capabilities this year, it may be a good time to address critical misconceptions and dangerous wrong turns that can harm the measurement journey, the most fundamental being that there is one single measurement solution which solves all marketing questions. As we shall see, it's only by connecting multiple methodologies, technologies and skill sets into one unified framework that we can answer today’s marketing challenges.
There are three commonly used approaches to answering the question of “What is my return on marketing investment?”:
- Marketing mix modelling (MMM)
- Multi-touch attribution modelling (MTA)
MMM remains the gold standard of ROI measurement. Now augmented by machine learning and far-improved data collection technologies in the marketing stack, MMM is both infinitely faster and easier to deploy and scale than its predecessors. Its strength over other methods lies in quantifying incremental sales from marketing dollars. But it also helps us decide where and how much to spend, i.e set budgets and optimise the marketing mix.
Multi-touch attribution (MTA) is used to “value” digital channels by apportioning sale/conversion credits to the various touchpoints in the consumer's journey. MTA is therefore a handy tool for in-channel optimisation at a very granular campaign level, albeit only for digital channels.
Experimentations are popular for testing micro-elements (A/B testing) in digital marketing and things like store promotions. It’s a bit tougher (hence underused) for strategic media ROI measurement, as control groups in the real world are difficult to manage 100%, plus it can be harder to measure any longer-term tail effects of media.
Debunking measurement misconceptions
As with all tools, each is best used for a specific task in specific circumstances.
However, there can be confusion about when to use which – or ideally how to marry multiple methods together – particularly as the modern marketing landscape continues to evolve.
Incrementality and attribution
Too often, these words are used interchangeably. In fact, they are completely different. If you want to understand/justify returns from your marketing efforts, it's all about incrementality. The best-in-class tool here is MMM. Digital attribution only measures the relative impact of digital channels on a specific conversion action. It does not tell you what would happen if you stopped investing in marketing or specific channels.
This does not devalue the role of digital attribution. It is an excellent tool for diagnosis and prescription for digital tactical planning, drilling down to the campaign and keyword group level (which MMM does not). Correctly used, MTA can lower CPAs by as much as 20%. Marketers need to be aware of when to use what:
- MMM outputs for strategic, holistic marketing plans and cross-offline, online or below-the-line optimisation
- MTA for micro digital decisions (with a higher frequency of measurement)
Death of attribution
Another buzzy myth is the imminent demise of digital attribution. Ideally, you need cookie data to track consumer digital pathways. Hence the misconception is, if the cookie dies, so does attribution. However, there are a couple of reasons why attribution may have a longer shelf life than expected.
Firstly, third party cookies are currently in circulation, with the official deadline for phase-out of 2024. Assuming no further extension, it still gives marketers a year's worth of measurement opportunity, although third party cookies are becoming less and less ubiquitous as we near this date.
More importantly, there are multiple solutions for digital attribution which are privacy-proof and future-proof. Key methodologies include using clean room technology (such as Google's Ads Data Hub) which provides access to user-level data in a way that is privacy-compliant. As a caveat, there are questions around how or if clean rooms can be useful, given all the “walled gardens”. There are also time-series regression-based “cookieless” solutions to measure attribution, which are simple and effective, and can be used across walled gardens.
A unified way forward
Each of the various measurement tools mentioned above has its role. However, the real power comes when you are able to statistically unify all these approaches to get to one answer, blending the positives of each separate method together. For example, MMM can answer the big macro questions and get a read on incrementality; MTA is key for micro-digital tactics; and experiments, while time and thought-intensive to set up, scale and maintain continuously, can give insight into very specific macro and micro challenges and feed the wider measurement model.
What does this mean for marketers? The best measurement frameworks stitch together MMM, digital attribution and experimentation in a statistical way to provide synergistic value between the three techniques. When done right, this yields an enormous and sustainable competitive advantage. Once built, delivering subsequent updated results with a degree of speed (ie MMM in days, not months) is a further leap in the marketing effectiveness journey.
If you are starting from scratch in your measurement journey, break it down into bite-size pieces and get started on each piece one by one, instead of everything, everywhere, all at once. If you already have these measurement pillars in place, unifying them and systematising for much quicker results will gain you significant competitive advantage.