Creating great market mix modelling is not a profitable side hustle and it’s really, really hard, which is why many agencies to date have steered clear of doing it. But with budgets tightening and marketers hunting hard proof of ROI and effectiveness, some are now attempting to build their own. That risks results skewing towards agencies’ own interests rather than marketers’, warn Atomic 212° National CEO Claire Fenner and Mutinex Co-Founder and CEO, Henry Innis. That is why the two firms have struck a deal.

Numbers game

Over the past 30 years, we’ve seen the media landscape become increasingly complicated. The rise and rise of digital media formats has driven a need for supporting measurement frameworks and the advertising industry has certainly responded to support clients in this need. Now, as we face the death of the cookie and increasing government scrutiny on privacy more generally it’s time for media agencies to stop marking their own homework and embrace market mix modelling with third parties.

Attribution crunch
Demand for market mix modelling is growing strongly. Why is that? There are a couple of factors at play. The first is that many customers are managing the shift away from multi-touch attribution (MTA) models that can no longer be supported without the cookies that hold them up. While monitoring the customer journey remains important for designing a great customer experience, touchpoints on the customer journey can no longer be used as a reliable stand in for the effectiveness or otherwise of marketing. Market mix modelling can help customers and agencies to remove the noise and uncertainty associated with MTA and see what’s effective in terms of revenue.

Money talks
This brings us to the second reason for the increased interest in market mix modelling. It’s a tough year for marketers. People are being asked to make savings where they can. Without a doubt, the most effective way to identify where savings can be made and incremental growth identified is with market mix modelling. Many customers also find that talking in terms of revenue makes it far easier to convince a finance team that marketing really is creating value.

Expensive mistakes
So where does that leave agencies in all of this? Media agencies play a crucial role in guiding customers through the measurement transition already underway and in helping to identify strategic change in times of pressure; media agencies aren’t just another supplier, they’re a trusted partner. It makes sense for agencies to cash in on this demand by setting up market mix modelling consultancies and building their own solutions, right? Wrong. Here’s why.

Firstly, creating great market mix modelling is not a profitable side hustle. Good modelling is hard. Really, really hard. While you may be able to hire the right people to set up a heritage style consultancy that builds bespoke models from the ground up for each customer, it’s not going to be cheap. So, you’re already charging clients an arm and a leg.

In addition, companies like Mutinex have already realised that Saas-based foundational models that can support fast data refresh and learn on the go will outperform static PowerPoint presentations where it really matters: helping customers to get to insights quickly and make decisions. That kind of modelling isn’t just powered by data scientists, but developers and product teams. So, are you a technology company now? You have to ask yourself if you have the chops to pull that off.

Conflicted data
But let’s say you do want to “build” or “buy” so to speak, should you? Media agencies are customer partners in investment. Agencies make recommendations and then implement investment strategies and the performance of the agency is judged partially on the success of those recommendations. It’s one thing to provide customers with the results of their campaigns, but should agencies be responsible for assessing the effectiveness of this work?

Agencies’ own modelling has a valuable role to play, but we also think there is significant value in agencies working with an independent third party to provide market mix modelling. In some cases, modelling is presented as “neutral” in and of itself. Who can argue with data right? But the reality is that all models depend on a set of assumptions. Assumptions can be misleading. And before you know it, your model is biased towards the best interests of the agency and not the customer. It’s just good ethical hygiene to avoid even the perception that the results of a market mix model could be biased in this way.

Marking, not mark-ups
That’s why Atomic 212° and Mutinex have partnered to provide independent market mix modelling services to the former’s customers. There’s a whole other article on why we have chosen each other as partners. The point is that as valued partners, it is the duty of agencies to help customers to the next phase of marketing measurement, but that should not come at the expense of either agency profitability or unbiased reporting. Marketing measurement is changing. Let’s take this as an opportunity to draw the line in the sand and say no more to agencies always marking their own homework.

Article originally published on Mi3.