Though They Can Be A Pain Point, Marketers Should Embrace Data Clean Rooms

Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Hugo Loriot, managing director at 55.

A couple of months ago a friend of mine, who works for a large brand, asked me whether to add Google’s clean room, the Ads Data Hub, to his priority list for the end of the year. “Of course you should,” I answered. “It’s not perfect, but it can give you insights your competitors don’t have. Not to mention you will have to work in such environments sooner or later.”

Last week, I had lunch with him, and I asked if his request was approved. “No,” he said. “When I brought it up, my media agency said I couldn’t because I don’t own my advertising data in Google’s Campaign Manager. My CTO added that it was not the company policy to use the Google Cloud Platform. My media team did not understand they would have to use SQL, and Google sent me an email to inform me they don’t have any beta-tester seats anyway.”

I nodded and took the check.

This scenario, though common, shouldn’t discourage brands from testing clean rooms. Knowing what to check before putting the Ads Data Hub or a Facebook/Amazon equivalent on the agenda is the first step in adopting this new technology.

A clean room is by definition a closed environment where you can specify and, in Google’s case, run custom queries against advertising data sets that live behind a walled garden. It unlocks insights that you cannot get by simply pulling a report from the user interface, but it only spits out aggregated data, for 50 or 100 users. Typical use cases include reach and frequency analyses on long time frames or against specific populations, attribution modeling on viewable impressions and advanced audience insights on first-party data.

Clean rooms are the byproduct of regulations such as the EU General Data Protection Regulation (GDPR) and California Consumer Privacy Act (CCPA) on one side and the cloud wars on the other side. Clean rooms have been created to give data-hungry marketers ammunition while keeping regulators happy. They have been designed as the intersection of the CMO and the CIO realms, using cloud computing for marketing use cases and letting the new generation of data analysts run SQL and Spark processes against CRM and media data files.

Created based on trade-offs between walled gardens and regulators, IT and marketing organizations, data clean rooms are either half full or half empty, which is why they are not spreading fast. But they will get traction, and it’s important to embrace them now, despite some of the hurdles.

The obstacles

Yes, they require ownership of the underlying media data, but data ownership is a requirement – not an option – for marketers in 2019.

Yes, they require data processing skills in specific cloud environments, but isn’t it the right opportunity to start the conversation between the CMO and CIO and to translate business needs in data-science friendly language?

Yes, they have been in limited access for years with few killer apps, but marketers who want to stay ahead of the curve should create their own use cases rather than wait for Google, Facebook and Amazon to push plug-and-play solutions to the whole market.

Getting started

Before embarking on using a clean room, brands should clearly define use cases and align with media partners on their feasibility. If you have not tested the Google Ads Data Hub already, that is probably a good place to start, with more support and a broader list of use cases.

Brands should prioritize advanced reach and frequency, ad sequencing and granular audience insights over CRM appending and complex online-to-offline analyses, which might not be available (yet). Don’t be afraid to invite everyone involved – partners, agencies and cross-disciplinary internal stakeholders – to a kick-off where the vision is set and it is conveyed why this matters to the organization. People will be more helpful if they are onboarded early and see the benefits.

Then, clarify who will actually create the query and run it. The Google Ads Data Hub has a self-serve interface and a sandbox, but brands may have to let Facebook and Amazon teams handle the brief and come back with a csv file storage address in Amazon Web Services. Brands must be clear on the need to scale, even if they are running a pilot.

Media teams and agencies must discuss how to best take action with the new insights. For example, if they get insights on the optimal sequence of posts and videos to drive sales, they should make sure to work on a messaging strategy with their creative partner. If they get insights about the impact of viewable ad frequency on sales, they can implement an ad experiment framework to statistically test the impact of increasing frequency or target viewability.

Finally, they need to get ready to iterate fast in this ever changing environment. Set up regular calls with partners to get updated on new features and adapt use cases accordingly.

As much as cloud for marketing and privacy by design are here to stay, data clean rooms will not go away. The sooner brands get acquainted with them, the better; all the hard nuts they will have to crack to make them work will have to be cracked anyway. It’s just more fun to do it on a product that is still kind of a secret and that only a handful of their peers have access to. They might be surprised by how much value they can get today.

Follow 55 (@55FiftyFive55) and AdExchanger (@adexchanger) on Twitter.

1 Comment

  1. To be exact: ADH (and other clear rooms) were created in response to Google redacting their IDs from DV360/GA logs, to allow advertisers to run consumer-level analysis (within the clear room)
    Google was forced to redact its IDs from logs due to the tougher regulatory environment around data privacy.

    Reply

Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>