Deconstructing common solutions for first-party data activation
As 2024 is starting, I want to shed some clarity (sanity?) on the topic of first-party data activation, which is very confusing for the whole AdTech industry — especially for marketers and agencies.
If you’re a brand marketer who understands that media activation based on third-party data is dead and that one of the most promising alternatives is activation based on first-party data (if not, you should check out this guide), you’re already ahead of the curve. But you’ve likely not arrived at your final destination yet. You’re now entering the maze of “first-party data activation solutions”. In this article, I will outline how popular solutions actually work and will do my best to demystify them so that marketers, agencies, and publishers can make an informed decision before choosing the right solution.
1) “Customer match”, the old school way
First-party data activation is nothing new and has been a targeting solution for a long time. Think Facebook Custom Audience (LINK) launched in 2013 or Google Customer Match (LINK) launched in 2015. These solutions enable you to target an audience in a given publisher ecosystem based on your own list of customers or prospects, using simple retargeting methods or advanced lookalikes. To achieve this, you send the list of users to the publishers, either in plain text or via pseudonymous identifiers (e.g. hashed emails).
This approach is totally fine if you find it acceptable to send your first-party data to media publishers and trust the quality of their lookalike targeting. But some Data Protection Officers (DPOs) are not quite happy with this concept (surprise, surprise), so other solutions have arisen in the meantime.
2) via a CDP
In the past few years, Customer Data Platforms (CDPs) have slowly but surely become an essential part of brands’ MarTech stacks. A CDP consolidates customer data that would otherwise be siloed across a CRM, a DMP, and other databases to create a 360° view of your customers. Additionally, it enables you to “activate” this consolidated first-party data. But how does it work? Essentially, it functions the same as the “old school” way described in section 1, but provides the convenience of connecting to the various “customer match” APIs out there in a one-stop shop.
Once again, this approach is totally fine if you are comfortable ceding your first-party data to media platforms. Some brands that buy a CDP get a veto on using the magic “first-party data activation” button from their DPO.
3) via an identity provider
Identity providers like LiveRamp or Epsilon also enable brands to activate their first-party data. To make this possible, you must first translate your customers’ personal identifiable information (PII) — for example email addresses — into a pseudonymous ID (e.g. RampID or CoreID). This pseudonymous ID is then sent directly to media publishers and DSPs. The identity provider acts as a neutral broker of first-party data, as it also performs the same translation job on the media publisher side.
One major pitfall here is audience reach, as this type of activation primarily focuses on retargeting and doesn’t offer audience extension based on lookalikes. Secondly, bear in mind that these pseudonymous IDs are considered personal data under GDPR and not anonymous data as I read or hear too frequently. So to put it simply, you must be comfortable with the fact that your customer personal data is shared with a multitude of third parties. To make things worse, the identity provider may also be monetizing your PII data to enrich its proprietary ID graph.
领英推荐
4) via a traditional data clean room
Cleanrooms are neutral and secure spaces where you can combine your first-party data with various partners like brands and publishers. Amongst the many use cases supported by cleanrooms, you’ll find first-party data activation. How does a cleanroom do it differently than a CDP? In many cases, it does not. When the cleanroom is used as an audience builder to enrich your first-party data with a data partner, the activation step is exactly the same as in a CDP, with the same advantages and disadvantages.
However, it gets interesting when the publisher is invited to the cleanroom as well. The cleanroom enables you to match your own first-party data against the publisher’s. In most cases, the publisher (and, less often, the advertiser) retrieves the list of matched customers from the cleanroom to create a tailor-made audience for you that you can activate.
This may look like an improvement compared to the previous solutions listed so far, since you no longer send all your proprietary data to the publisher, who instead can only access the overlapping customers. However, it’s not the perfect solution, as you still share some of your first-party data with the publisher, which triggers heavy legal contracting like data sharing agreements or joint controller agreements. Moreover, you have no control over how this seed audience will be treated (e.g. quality of the lookalike model, risk of re-use with a competing advertiser).
5) via Google PAIR
Google Publisher Advertiser Identity Reconciliation (PAIR) offers a more sophisticated approach to first-party data activation than traditional cleanrooms by adding layers of cryptography to make sure that neither the publisher nor you can learn who the matched customers are. However, it focuses entirely on retargeting and does not support lookalike activation, resulting in limited reach. In addition, PAIR complicates the already quite complex programmatic media buying workflows by introducing a new ID in the bidstream for each publisher.
6) with Decentriq data clean rooms
Decentriq offers a unique approach to first-party data activation, as it guarantees to you:
In the end, the choice of the right solution for you will depend on several factors, like your media mix, your aversion to sending your first-party data to third parties, your media budget, and your MarTech budget. If you’ve read this far, now you should be better equipped to navigate the maze and choose the right solution. Please get in touch if you are interested in option 6, obviously!
@thenunatakgroup. Innovation, Digital Strategy, Transformation, Change and Sustainability.
1 年As someone currently building up an identity solution for Switzerland (https://oneid.digital/) I'd like to get a better persona than the guy in the middle. Possible?
Helping ideas grow. Passionately. Purposefully. Beyond hype. Pushing boundaries to execute at the intersection of business, technology and autonomy.
1 年I like the idea of trying to compare these approaches. But don’t forget what informed consent, selective disclosure and sovereign control will and should do for the industry. I’ts not only about comapanies’ ‘owned’ data, but the future of how to ensure sustained sovereign control over the data as the first-person/holder. While data clean rooms provide unlinkability and great meanso of collaborating on data without revealing crown jewels, this is not the case with organisations that prevail in the data economy today. What you describe in regard to identity solutions is correct, when an IDP intermediary is providing the service, but not when you apply Self-Sovereign Identity (SSI) principles to your architectural paradigm. In my opinion we need to ensure that granular consent (and a regulatory and ethic framework) becomes a transparent and interoperable tool for the holder to execute control and agency across the data value chain. Whether the data is computed in enclave or not.
Vice President of Product @ InfoSum | Privacy Advocate | GTM Specialist | Product & Product Marketing Leader
1 年It could be argued that sharing data at any point in the process with another party within a clean room would negate the clean room title. I think we can both agree there. There are so many factors involved with a first-party data solution as most of the advertising and marketing industry is still clinging tightly onto their precious cookies. Pure play data clean rooms as we define them, are just true data clean rooms at their core built to protect the privacy of EVERYONE'S data. To me, that is the biggest caveat or difference between all the other solutions you mention. When working with first-party data it's not just your data you need to be cautious of or pay respect to. The whole point of data collaboration is to leverage other data with your own to improve performance/insights/etc. That is not a new concept. But collaborating with partners so that no one is in control no one is taking advantage and no one is in the dark about what is happening with their proprietary assets - that is the whole reason we are seeing a convergence of data clean rooms and data collaboration. Businesses who understand that will look to select the right DCR solution on scalability, functionality, partner network, ease of use, and cost.