Brand Measurement in a Post Third-Party Cookie World – A Q&A With Experts From the Research Committee
2020 kicked off with one of the most fundamental changes to the online advertising industry, when Google announced its goal of making third-party cookies obsolete by 2022. Where marketers had once relied on cookies for targeting, re-targeting, display advertising, and behavioural marketing in general, they now needed to change tact. And even though this has now been postponed until late 2023, the race is and continues to be on for the digital advertising and marketing ecosystem to find a solution that offers the ability to target ads in an effective and privacy-first way.
But amongst the whirlwind of the industry’s response to the end of cookies, how does brand measurement fit into the conversation? When it comes to a post third party-cookie world we must not only think about the available alternative solutions and what the future of attribution looks like but also consider brand measurement as we move into 2022 and beyond.
So, what will brand measurement look like in a post third-party cookie world? We asked members of IAB Europe’s Research committee to share their thoughts and expertise.
- Anita Caras, Sales Insights Director EMEA, Data Solutions, Yahoo
- Cate McVeigh, EU Measurement Leadership (Marketing Science), TikTok
- Duncan Southgate, Director, Media Domain, Insights Division, Kantar
Q1. On the brand measurement front specifically, what do we know is no longer going to work once Chrome deprecates cookies?
Anita Caras, Yahoo “This will impact some research providers more than others, depending on their data collection methods and whether they have a direct relationship with the panelists they utilise.
In order to run brand measurement studies a pixel is appended to the campaign ad assets, this recognises whether that user is an OnDevice panelist or not via the ad server call. If they are an OnDevice panelist they are served a survey, if not, the call is ignored. This approach is already being used and we are seeing data from all browsers (including Safari and Firefox) as no third-party cookies are used in the data match.”
Cate McVeigh,TikTok “Ongoing cookie deprecation across browsers has had an impact on how to execute brand measurement. The quality of experimentation will continue to get more difficult as approaches to how to execute brand measurement shift. With third-party cookie deprecation, there is a shift to more direct vendor integrations. This will lead to a trade-off between methodological rigour and flexibility for insights. There will need to be a practical acknowledgment that fewer players can execute randomised control trials well.”
Q2. What do we know is still going to work in a post third-party cookie world?
Duncan Southgate, Kantar – “As cookies are deprecated, direct integrations for the safe passing of deterministic passive exposure between publishers and research agency vendors will help maintain the quality of digital effectiveness measurement. These will rise in importance, but this will never quite manage to achieve 100% coverage because not all users will opt-in or be “matchable”. Smart probabilistic exposure approaches (where aggregated media inputs are mapped to claimed media consumption behaviours) will complement the passive exposure approaches, and be blended into new hybrid solutions. These probabilistic routines will themselves become increasingly sophisticated. Beyond the measurement of live campaigns, we also expect to see a continued rise in the popularity of controlled exposure “lab” approaches which enable brand lift measurement fully independent of cookie tracking.”
Anita Caras, Yahoo – “First-party cookies will continue to exist and work in the short to medium term. However, 3 years+ from now, I would not be surprised if the usage of first-party cookies will also come under scrutiny, as will the tools and techniques used to maintain a persistent identity. We are seeing signs of this with Apple’s recent iCloud Private Relay announcement.
What is clear is that those that have a direct relationship with their consumers, who can offer a clear value exchange (e.g. premium content, services, or rewards) for their consent will become increasingly valuable.
At Yahoo, for example, we have our own polling mechanics on our DSP to ask our users questions but we are looking to extend the capability further to enable us to continue to capture key brand measures amongst our millions of opted-in users across Europe.
I believe this change will also push many data providers to explore the use of metered panels, an approach used by TV in many European markets ie. explore the provision of a single source measurement panel as a proxy for consumer behaviour. On the positive, there is full transparency on the data capture, measurement, and monetisation. However, on the flip side, you have the constant challenge of balancing for panelist bias whilst managing retention, scale, and costs. The diversification of the online medium means that to be able to capture meaningful coverage of website and ad usage would command panels in the size of 10’s of thousands – which has hefty maintenance costs attached.
An alternative would be perhaps asking consenting panelists to undertake a specific task and/or visit specific sites where the ads reside (a research method known as forced exposure); this is a tactic we again use today in the absence of other workable solutions and is regularly deployed by other media. We work with research vendors who can deploy best practices such as having a 24-48 hour delay in questioning a respondent post-exposure in an attempt to mitigate inflated results which could be gained from requesting brand responses immediately after exposure.
Although this technique can provide brand measurement results across various formats, where this gets more challenging to measure is where there are discrete or niche targeting tactics employed that you are trying to replicate in a forced environment and hence should be evaluated alongside first-party and/or metered panel approaches.”
Cate McVeigh, TikTok – “Systems will continue to work, but they will need to have new integrations. There will be a need for managers to train measurement staff and non-technical staff who consume brand lift studies on how to judge quality experiments versus cheaper/faster/easier methods.”
Q3. What is still unknown or needs to be considered when it comes to brand measurement in a post third-party cookie world? And what needs to be tested?
Anita Caras, Yahoo “We already have browsers that have blocked the usage of third-party cookies and hence can test and learn from these areas as we prepare for the demise of the cookie on Chrome. To me, the biggest unknown is whether the scrutiny applied to the third-party cookie will be extended to first-party data tracking and measurement tactics and if so, in what timeframe – this in fact has already begun with Apple’s introduction of ITP; where the use of first-party cookies in a third-party context on Safari browsers has been limited to a 24-hour window, see here for further info. Looking beyond the technology changes we obviously also need to keep abreast of further legislative revisions and/or restrictions and the implications for measurement.
Whatever the changes ahead, we are moving to a world of accountable, consent-based relationships with our consumers. If we get this right, we have a much higher probability of providing accurate and accountable measurement solutions but it does mean that the pool of people that we measure against will get smaller.
Online has prided itself on being a 1:1 medium that can market to an individual and in return get analysis back on an individual basis, but this may have to change, with further data aggregation and cohorts being formed to align with privacy expectations. Perhaps we, as marketers, publishers, planners, or buyers, need to reflect on the fact that we have already weighed up the costs and benefits of these alternative measurement approaches, deployed them for decades, and accepted them as effective solutions for other media so why not online? What is clear to me is that now is the time to plan for every eventuality.”
Cate McVeigh, TikTok “For legacy solutions, there will be a number of things to consider such as scale, budget thresholds, experimental contamination, and possible bias that can’t be accounted for. New integration types such as server to server, clean rooms, and cohort level analysis should be tested.”
Q4. What is your main piece of advice to advertisers willing to measure the brand effectiveness of their digital campaigns in 2022 and beyond?
Duncan Southgate, Kantar “As the Journey song goes, “Don’t Stop Believin’”. The underlying exposure tracking capabilities are changing, but both single publisher and comprehensive cross-publisher digital brand lift measurement will remain feasible and accurate during this transition.”
Anita Caras, Yahoo “Get prepared and act now… If we are to take on board any learnings from the GDPR experience it should be act early and act now…we should be making sure we are focused on providing clarity back to consumers on how we are using their data and the value they are getting in return, in the hope of gaining consent, to fuel our measurement systems of the future.
Google has delayed the phase out of third-party cookies on Chrome until the second half of 2023 – this is not a time to take a big sigh of relief and carry on as usual, it means that the testing period is a little longer and we should use this opportunity to the full.
Cate McVeigh, TikTok “Statistical significance is not worth much if it’s not reproducible. Preparing for changes in partner systems and supporting necessary methodological shifts are key. Having a learning agenda will help you focus on new approaches that will continue to help drive your brand metrics.”