Marketing News South Africa

What's next after third-party cookies?

With the introduction and evolution of multi-touch attribution (MTA), the media industry has moved closer than ever to understanding the impact of each touchpoint on a customer's behaviour, from signing up for a newsletter to making a purchase. These attribution models are imperfect, but their intention has always been to give marketers a 360-degree view of the customer so they can make smarter marketing and media decisions, reducing waste and driving higher ROI.

Unfortunately, most MTA models were built leveraging third-party cookies. To increase user privacy, Google Chrome, which is the gatekeeper for 60% of the web’s traffic, has announced it will deprecate third-party cookies in 2023. Safari and Mozilla Firefox, which represent another 23% of browsing activity, have already deprecated third-party cookies, and Apple went one step further with the release of iOS 14.5, requiring users to opt in to have their data shared with apps and advertisers.

All told, it’s estimated that more than 90% of browsing behaviour may soon be anonymous to current attribution models..

This move represents a paradigm shift for everyone within the digital marketing ecosystem, especially for brands. As advertisers work to retool their models for targeting, attribution, and optimisation, they’ll lean heavily on first-party data and new technologies like data clean rooms – secure environments where first-party data can be shared across parties without risking data leakage or exposing PII, to manage that data and keep it secure.

All eyes on first-party data

To continue to measure ROI on campaigns and initiatives, marketers will need to adapt and devise new ways to track and attribute a conversion across consumer touchpoints. First-party data will be key to such efforts. Brands everywhere are preparing for this shift by increasing their focus on capturing first-party data from customers so they can “own” the customer relationship directly.

Some brands were already well positioned on this front. For example, businesses that require users to sign in to use their services or sell their products directly to consumers have a trove of user data they can leverage long after third-party cookies are gone. CPG companies on the other hand, which tend to sell their products through retailers, are at a disadvantage and are quickly trying to build their data sets ahead of these industry changes.

The future is in secure data sharing

Attribution models of the future will leverage first-party data to resolve customer identity, and then match that data with first-party data from partners to determine marketing spend efficacy. The exchange of data between partners will increasingly occur within data clean rooms.

Data exchange within clean rooms is already happening in the walled gardens of Google, Facebook and YouTube, and it will continue after third-party cookies go away. Within the YouTube environment, for example, brands can use information their customers have shared with them to target ads to customers or customer look-alikes, and to track conversions. Outside of the walled gardens, large platform publishers like NBCUniversal are using Snowflake to create data clean rooms to enable a similar process.

To replicate the clarity of cookie-based attribution, attribution marketers will likely need to leverage clean rooms as well, inviting agency and identity resolution partners to enrich first-party data within them. Snowflake presents a compelling option for marketers to build or engage in data clean rooms, since much of their first-party data already lives in the Snowflake Data Cloud and can easily and securely be commingled with other data sources.

The data governance and double-blinded privacy these data clean rooms provide may make marketers more comfortable sharing their first-party data with media partners, which in turn can enable better enrichment, better targeting, and higher return on advertising spend. Although constructing a data strategy that supports first-party attribution modelling is a short-term challenge, it may ultimately drive superior business results.

Let's do Biz