Blog
With the rise of e-commerce, online advertising and the ability of companies to implement detailed digital attribution at the customer level, media mix modelling (MMM) has lost popularity. Many next-generation marketers are unfamiliar with this tool, which is key to optimising a marketing program. But now there is a need to rethink these tools, especially in the context of increasing limitations in access to data collection.
Imagine you are the head of marketing for a consumer goods brand in the US in the 1980s whose products are distributed through third parties. Your main advertising channels are television, radio, direct mail, and print newspaper advertising. Other than using coupons, there is no good way to know which people saw your ad and who ended up making a purchase because of it.
However, the emergence of MMM has changed this paradigm. In the late 1980s and early 1990s, economic statisticians began to develop techniques for estimating the impact of different advertising channels on sales. This kind of modelling became a key tool in the marketing departments of major consumer brands. In the digital era, however, the whole situation has changed again.
With the advent of the internet and brands selling directly to consumers online, another change has occurred. Thanks to advertising on channels such as Facebook and Google, we can see exactly which customers have viewed our ads and whether or not they have converted. So instead of having to use a complex statistical model to estimate the impact of our ad spend, we can see the "conversion rate" and ROI right at the top of the dashboard of any digital advertising platform.
Still, it turns out that tracking digital advertising isn't the panacea we were all hoping for. Especially in the case of last-click attribution, the importance of channels that build awareness and are at the top of the advertising funnel (such as video ads on YouTube) is often overlooked. There is also the question of how to integrate online and offline channels and recognise the difference between correlation and causation.
All these challenges point to the need to rethink existing marketing analysis and adapt to the new reality. And here again we come to MMM.
Unlike measurement systems based on user identity, MMM does not rely on individualised transactional data. Instead, it uses aggregated data from different variables and channels, such as impression, cost and conversion, to examine marketing effects. This means that MMM is a truly holistic and robust system that measures offline and online activity in one place.
Thanks to innovations in machine learning, MMM has evolved significantly, enabling the efficient delivery of advertising information. This allows marketers to optimise campaigns according to the level of detail and speed required. This is why MMM is still effective even as the data ecosystem changes.
Although traditional marketers at large companies have been using MMM to measure brand-level marketing tactics for years, it is a tool that is still largely untapped by performance marketers. This is primarily due to a lack of understanding of what MMM can do, as well as a lack of options when building models.
It's time for marketers, including those born in the digital era, to evolve their existing marketing analytics. But that doesn't mean a complete rewrite of the marketing analytics playbook. Quite the opposite. In fact, there is an ideal, time-tested solution: MMM.
It's therefore high time to rediscover the importance of MMM and start harnessing its power in the new digital environment that is constantly changing and evolving. Whether you're a traditional marketer or a performance marketer, MMM can provide you with vital insights and help you optimise your marketing campaigns. Without it, your strategy may be incomplete. And that would be a shame.
Do you even have a question?
You can contact us directly
