by Guy Gobert-Jones, Managing Director of Digital Performance at Omnicom UK
How do we ensure that we are making the right decisions in-platform to deliver optimum outcomes? On the face of it, this seems like a simple question. But the answer is becoming increasingly complex.
It used to be that it was enough to watch your platform numbers go up. However, with platform data being increasingly modelled, and the platform features required to maximise privacy-safe measurement becoming more complex, teams are looking elsewhere to build a full view of performance.
What’s happening in the world of data-led decision making?
For a long time in the world of digital performance, we were taught that data-driven decision making was easy. Tinker with the tools at your fingertips (keywords, targetable audiences, or ad placements), and watch the metrics you’re focused on (normally conversions) go up in the platforms.
This is no longer the world we live in – and in fact, it probably never was. In a landscape where more platforms are controlling our historical tools themselves with capable (and, of course, AI- powered) marketing robots, and data is increasingly modelled and estimated, old platform metrics often fail to capture the total picture when taken in isolation.
Now, this is not to say that the latest spate of privacy-centric platform measurement functions, like CAPI, Enhanced Conversions and First-Party Mode, are barking up the wrong tree, or not as important for campaign performance as the platforms would make out. Bidding is now a game of platform algorithms. And the golden rule of platform algo is that the more data they have, the quicker they can learn how to hit the desired targets, and the better they function at doing those jobs.
All of the platform features mentioned above are fantastic at expanding the targetable data ecosystem for the platforms, enabling us to measure more and fundamentally drive more conversions. However, a future facing (and privacy-centric) in-platform measurement approach, while vital for maximising the efficiency of platform bidding algo, cannot be solely relied upon for platform decision making.
This has always been true for those big strategic decisions. There has been a general agreement for years now – use the platform data for your daily optimisations, but MMM and econometrics will always be your best friend for macro strategy and budget allocation. My question is, if we all agree that these advanced measurement techniques, underpinned by A/B & GEO testing, are best for the most important top-level decisions, then do we not also need to find a better way to integrate them into the 100 daily optimisation decisions that we make in-platform every day to guide towards that overall strategy.
Before you all shout, “We’d love to, but the scale and timeframes required for that big picture measurement are incompatible with the cadence of daily decision making”, I say, if the transformative power of AI and automation has taken campaign build times down from weeks to (sometimes) hours, and enabled an ecosystem shift from buying media based on target CPCs and CPMs to buying based on predicted life time values and propensity scores, then should it not also transform our ability to translate the big picture measurement that underpins our overall strategy into tangible daily optimisation decisions?
A future state of platform decision making
So, what does, or could, this new world look like? Predictably, I think it will be a balance of people and machine power. Clients are increasingly asking more specific questions of their advanced measurement models. Not just “how is my search performing”, or “how does my brand search compare to my generic”, but “how is my Performance Max showing up compared to my Demand Gen”. Not just “how is my social showing up?” but “what difference do creator led assets have on overall incremental value?”
As the technology for more agile MMM develops, I think we are going to start asking even more specific questions of these models more frequently. Questions like “what impact are my bidding strategies targets having on over all incremental value?”, and “what about my approach to audience targeting, and my keyword match types?”.
As these insights become increasingly scalable, technology will also progress to enable insight integration into our platform approaches; upscaling and downscaling investment based on where external measurement approaches identify maximum value, tweaking bid strategies, audience targets, and keyword coverage to maximise outcomes.
So, will we all ride off into the sunset of a robot utopia (at least as far as ad effectiveness and platform optimisation goes)? Not quite. Because there are challenges here as well as opportunities – ones that our platform specialists are best placed to understand and solve. I increasingly hear from clients that they are feeling ‘decision fatigue’ caused by data snow blindness.
This is where people and process are still the best tools in our arsenal. To borrow some language from one of our wonderful OMG agencies – they bring together the heart and the science. Which is why I think we are not only going to see more of this measurement data pushed directly into platforms for enhanced automated decision, but also better connection of insights from various sources across platform data, analytics data, geo tests, and econometric models into strategic decision making processes, that help our teams make the smartest optimisations day to day.
So maybe a human person, riding a robot horse into the sunset?
