3 Steps For Developing In-House IDs That Are Actually Usable
“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 Nancy Marzouk, CEO of MediaWallah. Marzouk will be speaking at Programmatic I/O, taking place in Las Vegas from May 23-25. You don’t want to miss it. Click here to register.
Several years ago, a host of brands decided to “in-house” their programmatic media buying in order to reduce middleman fees and increase transparency and control. But many soon realized they needed support managing technology and execution. So, rather than truly “in-housing” programmatic, what they really did was restructure their contracts.
The same thing is now happening with advertising data. As brands build out their first-party data strategies and work to assemble post-cookie capabilities, many are “in-housing” to remain in control of their data. A common strategy is for brands to create their own ID, which unifies their data and insights per unique profile.
To ensure success, brands need to accomplish a few important tasks. They need to develop a usable in-house ID, evaluate the data feeding into their ID, then fill in the blanks so their ID is usable in the real world.
Bringing data together
Unifying data to create a unique ID for each known customer and unknown visitor could be relatively simple for, say, a highly streamlined D2C retailer with just a handful of product lines. But it could be incredibly complex for a global company with many brands or layers of security around sensitive data.
Take a company like Chase Bank. With billions of highly sensitive data points, from account balances to social security numbers, the creation of an in-house ID is a complex process. Chase needs layers of encryption just to aggregate insights internally. To activate an ID, they’d need even more technology to create an entirely safe ID that’s disconnected from actual data.
In another example, such as P&G, the complexity lies less in data sensitivity than in logistics. With thousands of brands, unifying insights can mean integrating a variety of different technologies that don’t talk to each other, as well as managing permissions and coordinating entities that don’t always work together.
This early process requires C-level stakeholders to not only mandate coordination but also ensure that security and technology teams work effectively with marketing teams to provide safe access to insights. Companies can also launch a Center of Excellence that unifies technical and marketing teams with a common charter or set of goals.
Technically, companies must create a plan for data security and encryption, data aggregation and resolution, that can work across as much of the business as possible.
Evaluating the in-house ID
Once the difficult task of bringing data together has been accomplished, it’s time for teams to see what they have.
At this point, brands can run into the following issues:
- Old data: Ideally, an ID is refreshed with real-time data on a regular basis. Information that feeds IDs such as hashed emails, IP addresses and mobile IDs have a limited life span.
- Duplication or inaccuracies: If there are two emails for the same person, there could be two IDs associated. And not every customer will have the right information assigned to their ID.
- Data inconsistency: Some IDs might only include an email, while another only includes a physical address or digital ID.
- Lack of data for visitors: While many companies are adept at collecting information about their known customers, fewer are collecting relevant information from visitors.
Data needs to be reconciled or verified in order to ensure a company doesn’t have two IDs for a single person or IDs that are built on old or inconsistent data. To smooth this process along, many brands are creating an identity graph to knit different types of data together.
Filling in the blanks
The value of an identity graph is that it reveals what a company needs to do to have a more robust ID for use in advertising and marketing. After internal reconciliation happens, brands can look to third-party data sources to fill out an ID and reconcile issues. And they may add new capabilities like a persistent cookie to add insights.
Partners can also help brands with data enrichment. A brand often needs additional IDs to resolve information or to help for activation. And, of course, brands want to enrich IDs with third-party demo, location and other data to help with segmentation and targeting. For this process to work, brands need an activation key such as an identifier that can effectively stitch all of this data together accurately.
ID, meet world
Once out in the world, a brand will see how effective their ID is at reaching their audience at scale. In the next few years, there will likely be hundreds, if not thousands, of internal IDs that flood the market as brands and media companies try to reach audiences in new ways.
This will create complexity for technology providers and require interoperability to help everyone accomplish their goals. Brands should be open to testing with a wide variety of publishers and demand transparency that encourages everyone to work together.
Follow MediaWallah (@MediaWallahInc) and AdExchanger (@adexchanger) on Twitter.