What you need to know:

• Apple’s iOS update and the impending third party cookie cull are posing serious measurement challenges for brands – even e-comm pure-plays.
• Adore Beauty is turning to media mix modelling (MMM) and finding value in channels it previously discounted.
• Menulog is also using models to fill the gaps – and ride the price hikes – resulting from iOS impacts.

Foundation shift

“Dynamic media mix modelling is the future”, according to Michaela Michaut, head of acquisition & retention at Australian cosmetic pure-play, Adore Beauty.

Adore has quadrupled marketing spend over the last three years and grown massively as a result. But as cookies depreciate and Apple restricts tracking – hamstringing legacy digital measurement and attribution techniques – the ASX-listed firm and its shareholders needs to know which channels are performing.

As a result, Adore is pushing into media mix modelling (MMM). Michaut thinks MMM techniques and platforms will rapidly evolve as marketers scramble to find ways to understand which channels are performing without the crutch of cookies – with Apple’s iOS changes already starting to bite.

The firm’s recent MMM exercise “revealed some really excellent insights for our business that I think will help us to – particularly at that C suite level – be able to push some channels that traditionally from the last touch perspective looked to not be as strong … but have actually turned out to our most valuable,” Michaut told Mi3.

New order

Menulog CMO Simon Cheng agreed. The incoming changes from Google and in-play from Apple are already driving up digital marketing costs “significantly” while obscuring results, he said, leaving marketers more reliant on modelling as platforms and regulators upend the status quo.

While Menulog has seen higher Apple tracking opt-in rates than Australia’s – at best – low double-digit iOS 14.5 averages, “you lose the ability to track a pretty material proportion of your customers,” Cheng told Mi3.

“Naturally, that means you've got a smaller sample size to be able to collect data from. Your CPMs and your CPAs rise – and it’s a pretty material cost increase. You lose sight of what a large portion of your audience are doing; it's hard to optimise. And so I'm sure every business is starting to grapple with how they model their spend to be able to give a more accurate representation of what the world really looks like. We're grappling with that as well. And we have early models for how we can predict that.”

Cheng said Menulog doesn’t "officially" call its approach media mix modelling. “But we do both econometric modelling and attribution modelling. We do it on a less frequent basis than weekly, but we're constantly optimising the channel mix on the performance side of things to make sure we are getting best bang for buck,” he added.  

“I think there's going to be a much heavier reliance on modelling as cookies start going and all the various privacy changes come about.”

But because there is an element of guesswork in modelling, “those modelling exercises will have to be much more frequent”.

As such, Cheng thinks that “access to reliable information to inform good decisions” will be one of the biggest post-cookie marketing challenges.

Michaela Michaut and Simon Cheng were interviewed on post-cookie strategies for Mi3’s forthcoming first party data report, publishing mid-November, which provides a snapshot of marketer, publisher, media owner and agency strategies for a post-cookie, post privacy world.