clean rooms for beginners //

Enterprises have been forced to change their data strategy for customers in a big way. With the imminent deprecation and removal of third-party cookie and privacy and compliance at the forefront, they had no choice but to make significant changes.

Data Clean Rooms are a popular option.

What is a data clean room?

A data cleanroom is a collaborative space where two or more participants (brands or publishers, advertisers or groups within an organization) can come together and share or combine their first-party data.

They can then use each other’s data to gain insights under strict control. The collective data pool allows each participant to gain additional insights.

The data clean room is a place where each participant provides its own data. The DCR uses advanced algorithms to match data between participants and then adds additional attributes not previously available.

Each participant has more data and is able to run analytics, segmentation, and insights. Their marketers are now able to run more targeted activations — and, therefore, more efficient and successful — for example in paid media.

All of this collaboration and enhancement happens in a neutral, privacy-friendly environment. A contract governs what each participant is allowed to do and not with additional data. Controls also dictate how data is ingested, which matching rules are used and how data is activated. Here, security, governance, audit trail, encryption, and anonymization are all important.

The primary goal is to enhance datasets and collaborate. The data that has been enriched, if at all, will not be exported to the original source or participant. The contract between parties determines what each party can do with the combined data, within the boundaries of the clean room.

Data Privacy, and Compliance are the most important. All personally identifiable information is encrypted and hidden. PII is not accessible to any of the parties.

In the context of DCR, data exchanges and markets can also provide data enrichment capabilities. These are distinct and separate services. Some players will provide their DCR capabilities.

Data clean rooms: An example

The figure below illustrates how a firm that produces fast-moving consumer products (FMCG) might work with one of their large retailers.

The customer data for this FMCG consists of primarily demographic data (e.g. age group, location), and user preferences (e.g. favorite ice-cream). These data could have been collected in many ways, but let’s assume, for this example, that customers registered on the FMCG community website and provided the information.

The firm does not have transaction data in this case because it does not sell directly to consumers. However, their large retail partners do have transaction data including the date of purchase, amount spent, products purchased, etc. Retail partners also have results of campaigns on different social media platforms.

Both partners will benefit when these two decide that they want to work together in a clean data room. They can now segment more precisely and get new insights, as they have access to the additional attributes. Without this collaboration, neither partner would be able to achieve the same results.

Consider, for example, a case of paid media where there has traditionally been a high reliance on third-party information. In a DCR, both partners can use matching to improve targeting. Since the partners have access to the campaign results, they can determine if the same person has been targeted in multiple channels. They can then decide whether they want to minimize this overlap.

Dig deeper: Marketing use cases for data clean rooms

Data clean rooms: the challenges

DCR is no different. The following are some of the key challenges:

• Setting and agreeing upon the scope of data sharing. The data clean room is intended to be neutral but the rules are often set by the owner of the room.
• Governance and monitoring at a time of compliance
• Find the right partners who share the same DCR.
• You will need to use it with other technologies and tools.
• Last but not least, the technical challenges that come with integrating your stack. This includes data management, matching configurations, and integrating it into other applications.

Dig deeper: Evaluating data clean rooms for your organization

Data clean rooms are available in different types.

A data clean room may provide many different services, including data storage, identity matching bits, security, encryption and enrichment. You can therefore find data clean rooms with a variety of services on a wide market.

Ad hoc, players may partner to offer a comprehensive service. Finaly, some of them offer a vertical- or domain-specific offering that could be useful.

These players are grouped into five categories.

• Cleanrooms for data cleaning
• Data lakes/data warehouses
• Media companies and walled gardens
• Data onboarding vendors
• Customer data platforms

All of them have their own unique capabilities and differences. who controls the clean data room has a significant impact on governance.

Cleanrooms for data cleaning

There are many clean room specialists who specialize in this area. They may offer a variety of capabilities as independent players. This includes data enrichment through their data partners, in addition to the partners you have.

Most are smaller companies with limited market reach. Your potential partners are less likely to choose the same vendor. Negotiation may be necessary to get the right partners on the same platform.

Data lakes (DWH) or data warehouses

All leading DWH/data lakes vendors, such as Snowflake and Google, offer an optional service for data clean rooms. Some vendors offer a toolkit and require you or another firm to create a clean data room by using SQL, rules, table joins, stored procedures, etc. These providers extend their offering via a third party marketplace with additional partner tools.

If you already share the same platform with your partner, you might not have to physically move data. Be prepared to rely more on SQL and programming than visual interfaces.

Media companies and walled gardens

The term “data clean room” is not new, but walled gardens predate it. Google, Meta, and Amazon are the leaders in this field. These walled gardens allow you to ingest customer data and compare it with massive amounts of advertising data. Google, et. al. Google et al.

This is an additional option to the DWH offered by Google and Amazon. While still built on their DWH (e.g. BigQuery for Google), the partner data is limited to advertising data from that walled-garden.

Some large media companies offer a clean room service in addition to these walled garden offerings. These offerings are specific to the media destinations of these large players.

You may be surprised to find familiar technology under the covers. Disney’s data cleanroom was developed in collaboration with DCR vendors Habu, Infosum and Snowflake. NBCUniversal’s Audience Insights hub also works with Snowflake.

Data onboarding vendors

Data Clean Rooms are now offered by several data onboarding providers. These vendors offer additional useful capabilities such as identity resolution, access to their data marketplace and the ability to leverage data not only from your partner but also their own network.

This alternative is useful for matching data sets between partners and enriching data from the first party with data from second and third parties. Their activation capabilities may be limited.

Bonus category: Customer data platforms (CDPs)

Only a handful of CDPs like Adobe and Blueconic offer private DCR functionality to their licensees. This also means your partner has to use the same CDP. The network effects are therefore limited. Your first-party data will remain in your CDP, so you won’t have to move it anywhere else.

DCRs are a powerful tool for activating data.

Data clean rooms have become a popular way to increase the return on investment for your customer data. There are many options available, but you should consider the following key factors when choosing:

• The options listed above have several overlaps. Vendors from these options will often work together to offer their data clean rooms. As an example, a Snowflake data clean room could be used with a vendor in its marketplace or another vendor’s offering. They may both be similar, but they are offered by different vendors.
• You and your partner are required to make the data in the same clean room available. It may restrict your options for collaboration.

It is therefore not unusual for enterprises to use multiple clean room data offerings, depending on the use case and partner profile. A savvy business will always keep all options available.

The post Data Clean Rooms: A Beginner’s Guide first appeared on MarTech.