Identity resolution technologies are tools and platforms that help businesses connect and unify customer data across various touchpoints and devices to create a single, comprehensive view of each customer. 

They have several important aspects:

Data collection and integration: These technologies gather data from multiple sources, including online and offline channels, devices, and platforms.

Identity graph: A core component is the identity graph, which is a database that connects a customer’s profile to all known personally identifiable information (PII) for that specific consumer. Let us explain the identity graph in simple words. 

Imagine you have a big magical sticker book about you. This book is your identity graph. It has pages for all the different things about you, and each page represents a different part of your life. The stickers are pieces of information about you. They show what you like, what you do, and where you travel. Some stickers connect to others with colourful strings. These connections show how different parts of your life are related. Examples of stickers are your name and birthday, your favourite toys, books, songs and games. Companies who make things for children – like toys and games – look at something like these sticker books. It helps them understand what kids like you enjoy, so they can make better things for you. 

Data matching and linking: Identity resolution platforms use various methods to match and link different data points to a single customer profile. 

What is data matching? 

Imagine you have a big box of toy animals. Some are lions, some are elephants, and some are giraffes. Data matching is like trying to find all the lions and put them together in one group. 

What is data linking? 

Let’s say you have another box with toy trees and grass. Data linking is like connecting the animals to the right habitats. You are making connection between different pieces of information. For example, you could connect the giraffes to the tall trees they like to eat from.

Persistent ID creation: Companies try to create a unique, persistent identifier for each customer that remains consistent across channels and over time. 

What are identifiers in simple words? 

Imagine you’re in a big classroom full of students. Each student has a name, but sometimes two students might have the same name. This could be confusing. To make sure everyone is clearly identified, the teacher gives each student a special number – which is unique. No two students have the same number. 

An identifier for customers works in a similar way. It is like a special unique tag or code. No two customers have the same identifier!

Data onboarding: This involves moving offline data to an online environment to connect offline customer records with online users. 

Let me explain data onboarding in simple words. 

Imagine you have a big box of toy cars in your room. These cars are like your offline data – they’re not connected to anything, just sitting there in the box. Now, you want to play with these cars in a fun online racing game on your tablet. But the game doesn’t know about your toy cars yet! Data onboarding is like carefully taking each of your toy cars out of the box, and putting them into the computer game while creating a digital version of each car. Once all the cars are online in the game, you can race them and have fun! 

This is just like what companies do with important information. They take data that’s stored in old ways (on paper or old computer systems) and move them to new, online systems where it is easier to use, share, and keep safe.

Privacy compliance: Modern identity resolution technologies ensure compliance with privacy regulations like the European GDPR and US CCPA. We will discuss privacy regulations in a later article.

APIs and integrations: Identity resolution technologies often contain APIs for integration with other marketing and customer experience platforms. 

Imagine you are playing with your favourite toy kitchen set. You have a little plastic stove, some pots and pans, and even some pretend food. How do you use this toy kitchen? You put a pot on the stove. You turn the knob to ‘cook’ the food. You take the ‘cooked’ food out of the pot. In this example, the know on the stove is like an API. It’s a simple way for you to tell the stove what to do without needing to know how the stove actually works inside. 

An API stands for Application Programming Interface. It is a simple way for computer programs to talk to each other and share information or do tasks.

Deterministic and probabilistic matching: Identity resolution technologies may use deterministic matching (based on known identifiers) or probabilistic matching (using algorithms to predict likely matches). Deterministic matching is like finding exact pairs of socks. You look for socks that match perfectly fine in every way. They must have the same colour, pattern, size, and brand. If everything matches exactly, you’re sure it’s a pair. You can say: “These socks are definitely a match because they’re exactly the same!” 

Probabilistic matching is more like finding socks that are probably a pair. You look for socks that are very similar but might not be exactly the same. Maybe the colour is a little faded on one because your Mum washed it accidentally in a washing machine at too high a temperature. You decide how likely it is that they’re a pair based on how similar they are. You can say: “These socks are probably a match because they look very similar.”

Real-time processing: Many platforms offer real-time identity resolution capabilities to support immediate personalization and targeting. 

Real-time means everything happens super-fast. When you do something, the computer reacts right away. The computer is always ready to receive new information. Unlike when you are waiting for a video to load, real-time processing doesn’t make you wait. 

Machine learning (ML) and AI: Advanced identity resolution systems use ML algorithms to improve matching accuracy and identify non-obvious relationships between data.

We interviewed Anthony Scriffignano, PhD for our podcast, “AI Snacks With Romy and Roby.” AI has over 102 patents, many of them on behalf of optimising identity resolution technologies. Episode 17 is with Anthony is fully dedicated to identity resolution.