Identity graph

What is identity graph

In simple terms, an identity graph is a database collecting and connecting different pieces of information about individuals to create a complete picture of their online behaviors and interests. These pieces of data are called identifiers, and they can, for instance, be contact details like email, phone number, home address, and postal code, as well as device ID or cookie ID. As new information comes in, the identity graph updates itself in real time. The goal is to create the most complete and up-to-date picture of a number of people.

What’s vital is that within an identity graph, customer profiles can be categorized into two types:

• Authenticated profiles are profiles that get their data thanks to the fact that the user logged in (using emails or usernames). In consequence, it’s possible to track users across devices and channels reliably;
• Non-authenticated profiles rely on cookies or device IDs but are temporary and less accurate than authenticated profiles.

Since massive plans of phasing out third-party cookies in Chrome (that were finally put on hold) and their declining reputation in the industry, publishers should rather create identity graphs based on authenticated profiles.

Why publishers should create identity graphs

Here are reasons why, as a digital content creator, you should consider making an identity graph:

• An identity graph based on authenticated profiles is an excellent solution in a time of many strict privacy laws. It’s because first-party data respects user privacy as it’s data you gather yourself;
• With an identity graph, you can get a better grasp of your audience and their preferences, but also behaviors and basic information like age and gender, location, and more. In fact, people can use different devices (like phones, tablets, and laptops) and different accounts (like emails or social media) when displaying your content, and with the identity graph, you can see the whole picture;
• Identity graphs help turn anonymous website visitors into profiles with vital pieces of information about individuals, which can later be segmented;
• Due to better understanding your audience, your content can become more precious to specific advertisers, whose main aim is to reach well-targeted audiences.

Step-by-step identity graph guide for digital publishers

1. The first step is collecting data in a single place, creating a storage system by using, for instance, a cloud-based data warehouse (like Google BigQuery that can complement solutions we disclose in the next paragraph). After that, it’s time for data normalization, in other words, fixing errors, ensuring everything matches, and putting everything into a standard format so it’s easy to work with;
2. The second step is all about creating a complete and accurate profile of the individual, thanks to the method of joining fragmented, scattered data. ID graphs rely on two approaches to match data:
   • Deterministic matching (explicit matching) relies on identifiers like log-in details; it’s like recognizing someone by their ID card – you know for sure it’s them. Many publishers decide to add the option of creating a free account, among others, to collect data for more accurate ads,
   • Probabilistic matching (implicit matching), on the other hand, uses data like screen resolution, location, IP address, and more. The most crucial fact is that this kind of matching isn’t as accurate as the explicit one;
3. The last step is building the graph! Each person is a basic unit of a data structure, and edges are lines between nodes that show how they’re connected, like:
   • Which devices they use (phone, laptop, etc.),
   • What actions they take, and more.

Basically, a graph is a web of nodes and lines that help you see patterns. However, you have to remember that after creating an identity graph, it must be regularly checked and updated.

Tools that might help with identity graphs

There are many companies and tools that can help you create an identity graph. Some of them are listed below:

• “Panorama Graph” by Lotame – an identity resolution engine that helps to get a 360 overview of users with the help of machine learning models. You can request a demo if you’re interested in this solution;
• LiveRamp’s “Identity Resolution” offers an identity infrastructure that enables businesses to collect and integrate data from various sources and create a comprehensive view of each customer. Contact the sales team for more information on pricing;
• “Identity” from InfoSum enables businesses to maximize match rates and enrich audience profiles, offering targeting accuracy. As in the case of solutions mentioned earlier, it’s best to contact the company’s team to get all the information you need.

All things privacy

Following the steps described above helps publishers build a reliable identity graph that makes it easier to identify audiences, tailor content, and grow revenue, all while ensuring they follow the latest privacy laws. If you wish to know more about privacy-friendly options in the realm of digital advertising, we recommend getting familiar with the information described in our article on contextual targeting.