By Nancy Marzouk, CEO and Founder at MediaWallah
Every few months a new term gains popularity and people start tossing it around. This phenomenon has become so ingrained in the ad tech industry that we have all come to expect a new “WTF is…” article to come out as soon as any new term arises. Terms like “programmatic advertising” or “supply side targeting” aren’t particularly intuitive to the average person and so people feel comfortable asking what they mean. However, in other cases, a new term arises that does sound intuitive and people start making assumptions.
Identity resolution is one such term. Clearly it means “resolving” something that’s related to someone’s identity. Most people assume this means bringing data together about an individual. At a basic level, that assumption is correct but it can get companies into trouble. Identity resolution actually refers to a variety of actions in order to resolve data into a unified record and companies won’t have a complete view of their audiences unless they understand the full picture of identity resolution.
Owner-Centric Resolution Is Half The Picture
A brand or publisher with a lot of disparate data is going to want to compare two data sets, deduplicate similar data sets or resolve data from two different places in order to have a “golden record” of an individual. I tend to call that “owner-centric” identity resolution because it’s focused on bringing owned data sets together or matching another data set to the owned data. For owner-centric identity resolution, a company is likely going to use their CDP or a data cloud partner like AWS, Google or Snowflake. This is the first and most basic identity resolution step, basically getting the house in order.
It’s important to know how owner-centric identity resolution works because it’s fundamentally limited by its mechanics. To create a match between two records, there needs to be a common key. For example, a first party cookie that can match two different site visits or a common email address to match a purchase to an email click. These matches can be done using a query within the CDP or cloud environment and can be done securely so that everything is encrypted.
Owner-centric resolution is a great place to start but it has inherent limitations. First, the requirement of a common “match-key” leads to low match rates. And second, the record is limited to whatever data is available in those specific data sets, which might not be enough for the company’s marketing, advertising or analytics requirements.
Third-Party Identity Resolution Fills In the Gaps
There are many reasons why companies would need to go beyond matching two data sets with a common key and turn to a third party match approach. First, some data sets simply don’t have a common key. Second, two data sets don’t always resolve every identity – there are often a lot of data that isn’t matched (hence the low match rates.) And of course, companies could want to enhance the insights they have for an individual record with additional information, either to complete their records or to get more depth to their 360 degree view of the customer.
Not enough brands are aware of this other part of identity resolution, which relies on matching logic and an identity graph or identity spine to connect data that doesn’t have a common key. Often, companies have what they think are unique or complete records, which are incorrect due to a lack of third party matching. One common example is to have a different profile for the same person who has two different email addresses. A third party could use a variety of device IDs, IP addresses and other shared info to correctly combine records and know enough not to combine two people who happened to sit near each other at work or at a coffee shop with a shared wifi. Third party identity resolution is key to matching data at scale and unlocking the value of all of the data across those different profiles. Having the ability to connect multiple ids for individuals allows for better addressability and more thorough analytics. What’s more, this kind of data matching can’t just happen anywhere. Privacy requirements dictate that data be kept completely secure throughout the process, which is why more companies are including data clean rooms in their identity strategies.
Preparing for Identity Transformation
The amount of data available to companies today is unfathomable compared to a few short years ago and is a fraction of what will be available in the future. In 2023, the world is expected to create 120 zettabytes of data compared to 97 in 2022 from a base of only 2 in 2010, just twelve years ago. The data matching that companies will need access to in the midst of this data explosion will need the depth and flexibility to go well beyond the current owned-data capabilities within a typical CDP or data cloud.
In the old days, third party cookies served as a sort of universal key for advertisers across their media buying and marketing. Today, many companies are just starting to get their first-party data in shape for use at scale and owner-centric resolution provides a solid base to start that journey. However, to get the scale once enjoyed with third-party cookies and to increase accuracy, brands will need to amplify the effectiveness of their data with broader efforts. Any company that is hoping for true identity transformation, where their marketing and advertising is powered by 360 customer records that can work across marketing and advertising, needs to include third party identity resolution.