EXECUTIVE SUMMARY
Before we get into the death of the cookie, it is important to look at the history of the third-party cookie, their use in the current programmatic ecosystem, and the reasons for them disappearing. Let us start by understanding what a third-party cookie is. A third-party cookie is a piece of code that lives on your web browser, and acts as an identifier to you and your browsing history.
Since the inception of digital advertising, the third-party cookie has been used by publishers and buyers to show relevant ads to users, and measure the effectiveness of the ads. However, due to a number of issues in the last few years around data privacy and transparency, browsers like Safari and Firefox have removed them. These browsers make up roughly 20% of share globally, so publishers and advertisers were still able to operate using the third-party cookie, knowing they could still effectively track the majority of users across the globe using this method.
The reason this is now such a hot topic for the industry, is that Chrome (Google) announced at the start of this year that they too will be removing third-party cookies from their browser, sometime in 2022. Chrome makes up approximately two-thirds of the browser share worldwide, meaning it effectively signals the end of the third-party cookie for advertising purposes going forward.
For this reason, it is important to start thinking now about how to tackle the new digital world minus the third-party cookie. Below are 10 things to start thinking about today and put into action.
GENERAL EARLY RECOMMENDATIONS
The good news is that third-party cookies have been going away for some time, and the digital world has not “crumbled”. Chrome is the last piece of the puzzle, before we move to a new world not relying so heavily on an old and somewhat inefficient mode of targeting and measurement. The first thing to do is not just wait and see how this plays out, get on the front foot and talk to your programmatic partners and publishers about their views on a cookieless world. Why are they suited better than others? Do they have a robust first-party data offering? Understand that it is not about targeting, it is about measurement. Look for partners that have already been exploring what advertising means in a cookieless world, specifically insights from Safari and Firefox activity. This is a great starting point for brands and agencies to understand which solutions are ahead of the curve.
GET COMFORTABLE WITH DIFFERENT MEANS OF TARGETING AND MEASUREMENT
As we move away from third-party cookies, a number of the various targeting and measurement strategies we have been so accustomed to will be gone. It will be imperative to the digital industry that we start getting comfortable about judging the success of our campaigns based on different metrics. This is a challenge for a number of folk in the digital world, because this is the only method of targeting and measurement they have worked with. Things like frequency capping, retargeting, post view attribution looks very different without a third-party cookie present. Think about developing new contextual targeting strategies, explore methods like incrementality tests or survey data to determine actual results of your ad campaigns. An example might be an advertiser not just looking at different contextual topics and formats, but also using their first-party data to understand what content interests their own customers have. This can assist in building more contextual strategies in the future.
CHECK IN WITH YOUR DATA MANAGEMENT PLATFORM (DMP)
If you currently have or are considering DMP, engage with your vendor(s) to understand how they are adjusting to the upcoming reality. Enquire about the sources of the data and how much of the data they have is deterministic or probabilistic. Deterministic data is directly declared by consumers or explicitly validated by the data provider, whereas probabilistic data is based on statistical samples and extrapolations. As a result, probabilistic data is usually less accurate than deterministic. You should also understand how much of the data comes directly from first-party data sources of participating suppliers, instead of aggregated. Aggregated and probabilistic data is still very useful but it is important to have a clear understanding of the generalisations made and how they might impact the effectiveness of your media buying. The ongoing developments are likely to decrease the amount of deterministic data available, replacing it with probabilistic projections.
CONSENT IS KING
In a context of having to ask for consent for personalised ads in your website or app, it is easy to just tick the box of having compliance with applicable consent mechanisms. However, it is far more powerful and effective to wrap the compliance with a very clear value proposition for gaining consent from your users. Some ideas to achieve this would be to provide easy to understand and compelling descriptions of how the consent and data help your users to get a much richer and satisfying ad-supported experience, as well as explaining how the data is secured to prevent undue uses or tracking beyond the specific purposes outlined in your consent request. Again, highlighting the value that will be given back to the user in exchange for their consent should be the primary goal, not simply achieving compliance.
THINK MOBILE FIRST AND CONSIDER HOUSEHOLD TARGETING
BUYING AROUND THE WALLED GARDEN
Media buyers will have to shift their targeting towards user data using user profile, opt-in or data gathering via login. A survey in US and Europe conducted by Sizmek in 2018, “walled garden” such as Google, Facebook, makes it difficult for marketers to evaluate data as walled gardens do not provide enough data for marketers to track the effectiveness of the campaigns. Therefore it is advisable for publishers to start to build their own version of walled garden and collect user data leveraging first party data. Agencies and brands will need to work with data providers leveraging combined information such as logged in information, purchase information, mobile device usage and location, together with publisher’s first-party data to craft out a targeting strategy.
MOVING BUDGETS TOWARDS OVER-THE-TOP (OTT)
In an OTT environment such as Viu, Hulu, Netflix and iQiYi where cookies do not exist, streaming providers make use of information such as device ID and content interest to build the audience targeting. With OTT inventory, buyers will be able to target users from an audience segment. Beyond gender, age, and location information, data such as duration that they watch, the type of message or ads users interact with and who in the family is watching the content can help marketers target specific audiences. Metrics such as video completion rate, interaction rate, attribution rate and visit rate could help marketers to measure ad effectiveness. Marketers can also leverage data like what TV content the user watches, time of the day, device type (mobile or tv), to target the right audience who have intent to purchase the product.
FOCUSING ON CONTEXT AND FORMATS
Marketers can look into contextual targeting to get insights about consumer information such as how often their users visit the page and user behaviour. Brands can work with publishers to target their campaigns based on the segments that the content being categorized without the use of cookies. Marketers who focus on high-impact formats such as skinner, video and branded content will have higher potential to outperform other brands to catch the consumer behaviour.
THINK DATA SAFETY AS AN EXTENSION OF BRAND SAFETY
In the complex world of digital advertising, brand safety and ad fraud are prevalent issues.
But as the industry evolves, there is a new pillar for marketers who want to build and protect their brand: data safety.
Data safety is all about ensuring that you are using safe data, not toxic data, for all advertising. By definition, safe data is any data collected or used with freely given and informed consent from the consumer, in compliance with all data privacy laws. While data privacy regulation in Southeast Asia and India has not yet reached the requirements expressed by the Europe Union’s General Data Protection Regulation (GDPR) or California’s California Consumer Privacy Act (CCPA), these can still be applicable to the significant amount of expats and tourists in the region. Meanwhile, there is rapidly evolving legislation in key markets like Singapore, India, Indonesia and Malaysia that will tend to follow the examples already enforced in advanced markets.
Data safety is a concept that ensures that the data marketers use for advertising is safe from reputational, legal and financial risk, and collected with trusted and traceable user consent.
BRINGING BACK BIG(GER) DATA MODELS
Oftentimes, when folks think of data-driven media planning, what first comes to mind is direct media addressability. This is especially true in adtech circles as the majority of discussions center around impressions traded within a programmatic auction. As such, the usage of more representative, population-level data sets (e.g. from syndicated research) upstream during the strategy phase (or even pre-campaign brief) is an area that can often be neglected by programmatic practitioners.
However, as we barrel towards a world without third-party cookies and a compromised mobile device ID, there is a real opportunity to merge these spaces and leverage traditionally upstream, aggregated data sets to augment and inform the deployment of downstream programmatic-era data sets in new ways. In doing this, we connect different “levels” of data in an unbroken thread. With the paradigm around the way we are able to collect and deploy user data moving forward shifting to one built on consent, it is important to keep in mind that the more representative, population level data sets will not be disrupted.
These data sets are representative of larger audience groups and while slower moving and less programmable, have inherently lower bias than user-level programmatic data sets. What we can do is use the former to create a strategic “sandbox” within which we can deploy event level data whilst mitigating the risk of bias. Just the same declared user level data can be used to “calibrate” syndicated research to fit a specific campaign use-case.
GETTING OUR HANDS DIRTY IN DATA CLEAN ROOMS
With the future of the industry being built around first-party data, user consent and privacy, the role of data clean rooms in marketing will become increasingly important. By definition, a data clean room is a space allowing multiple parties to bring their data together for analysis under defined rules to ensure user privacy and compliance. These rules generally control for the type of data that can be brought into the clean room, how that data can be joined to other data in the clean room, the type of analysis each party can perform and what data can exit the clean room. With open access to log-level data with non-redacted user ID fields becoming increasingly limited, data clean rooms allow marketers to continue performing advanced analysis for measurement, attribution and audience segmentation, albeit at an aggregated cohort level.
Within the industry, the most prevalent data clean room solution is Google’s Ads Data Hub, though solutions are also being developed by Amazon and Facebook. Outside of the walled gardens, cloud data platforms such as Infosum and Snowflake are offering decentralised or distributed clean rooms enabling data to be joined by multiple parties to create a privacy safe data co-ops.
THE FORECAST FOR TOMORROW’S AD STACK IS CLOUD(Y)
With all that we have discussed, it is clear that the traditional marketing stack of yesterday will need to evolve to account for a user consent driven world without third-party cookies. One trend that we believe will continue to grow is the increasing role of cloud-based data management systems that we have historically associated more with IT or Business Intelligence teams.
Integrations between ad platform and cloud applications are already taking place. As an example, marketer’s now have the ability to directly import and export data from Google Analytics to Google Cloud Platform via BigQuery. An even more direct example is that Ads Data Hub is built on top of BigQuery. Cloud-based data management applications will be integral in helping marketers store, analyse, model and visualise different types of data from ads-based data to Customer Relationship Management (CRM) data and research studies. This in turn will enable marketers to glean more holistic insights with which to make marketing decisions.
While extracting and manipulating data within cloud-based systems today generally requires a working knowledge of Structured Query Language (SQL), it is likely that low-code alternatives and options will appear over time, allowing more advertisers and suppliers alike to leverage these systems.
Source: iabseaindia.com
