How to jump off the cookie cliff without a serviceable parachute next year

The clock is ticking. Over the next twelve to eighteen months, we’re going to see the biggest shakedown in digital marketing since GDPR.

We’ve been talking about the demise of third-party cookies since before the global pandemic and even Brexit. But after years of incremental change and delays, are brands ready to finally jump off the cookie cliff? 

Some are nearly there having implemented most of the mission-critical solutions required, and the last few changes are on the roadmap. Others are still in an exploratory phase. They have not yet established a robust way to test and embrace a multiplicity of new solutions – probabilistic, contextual, IDs, first-party data, and external third-party data, all coming together in a pea soup of future measurement solutions.

Some of these solutions will deliver, some won’t, and the methodology each brand should rely on depends on the level of capability, resources, and guidance that they have access to. The unique needs of each advertiser based on their size and maturity plays a role as well. 

Whatever their size or situation, all brands will need a layered approach to investment measurement, knitting together a tapestry of solutions that can work together to create a relatively accurate picture of success. ISBA’s Origin framework provides a useful starting point. 

Marketers are increasingly calling this approach triangulation. Testing and re-testing across a range of measurement mechanisms to hone in on measurement truth as far as possible.

Table stakes: invest in marketing mix modelling (MMM)

Whether you’re a heritage brand or a start-up, you’re going to need to invest in MMM. 

Look at it as the foundation of your measurement pyramid. As the third-party cookie disappears, it is going to be essential to cross-check the data that will come out of a range of relatively untried new modelling sandboxes proffered by tech platforms and sense-check it against sales and other metrics.

Think long-term. Even the best marketing mix model may be relatively inflexible, but over time it can provide reliable, consistent data, and an independent and objective assessment of marketing effectiveness. It allows advertisers to verify the results provided by different platforms.

Independently curated MMM provides a more holistic view as it takes into consideration the impact of marketing activities across various channels and touchpoints, both online and offline. However, some models only get updated quarterly and sometimes six monthly. So insights can be pretty basic such as performance per channel, but not able to offer much insight between prospecting versus retargeting, or between specific audiences, tactics or creatives.

However, brands have more options than ever for MMM across a range of new opportunities such as Meta’s Robyn and Google Lightweight MMM. Every brand should ensure they have a strong marketing mix model in place as one of the key points of their new measurement triangulation. 

Plan first-party data strategy to reimagine attribution 

Google announced in September that it had switched on its APIs for Privacy Sandbox and while a proportion (3%) of Chrome users will remain unaffected so that Google can carry out some A/B testing, all users will have access in a matter of months.

As third-party cookies fade, cookie-based attribution modelling will become weaker. First-party data is only available via sites with logins and subscriptions, or from sources such as telcos. 

Using in-house first-party customer data will be critical, and brands need to plan now how they will collect and make the most of it to power better targeting (think more accurate lookalike audiences) as well as measurement. Meta’s Conversion API or CAPI allows brands to upload web, app, and offline data to enable ID conversion modelling to provide analytics based on activity within its platforms.

Google's Privacy Sandbox introduces alternative probabilistic attribution and targeting solutions that advertisers need to understand and incorporate into their strategies. But success again depends on the amount and quality of data available. 

Testing and experimenting with conversion modelling alongside other measurement approaches can lead to the discovery of new and innovative strategies that might not have been apparent or obvious initially. For example, we found that viewability was a better predictor of incrementality than modelled conversions for one brand’s retargeting tactics. This was a clear case where triangulating measurement across marketing mix modelling, media optimisation KPIs, and attributed conversions led the brand to a much clearer picture of incrementality and better performance outcomes. 

More and more, brands should be finding these kinds of insights as they move to new, future-proofed measurement approaches.

Experiment with a data-driven approach

The old ways of simplified cookie-based attribution will soon be gone. No single measurement method can provide a complete picture of the effectiveness of campaigns. Experimenting with new solutions is the path forward. Research shows how well a data-driven approach pays off: advertisers with a culture of experimentation see improvements of 20% to 40% in spending efficiency and 10% increase in advertising effectiveness.

Brands of all sizes now need to ‘triangulate’ between MMM, new forms of attribution modelling and lift testing, because their digital advertising toes are on the cookie cliff edge. 

Betsy Ray is the director of marketing analytics at Kepler EMEA.