Principles of Identity Resolution

Identity Resolution
You can deliver a more personalized brand experience with the correct identity resolution platform. And it can be done without compromising data privacy or security. Identity resolution is recognizing your customers in real time by matching the different IDs stored across all your digital channels, systems and data sources. It’s a vital part of your customer data strategy and is the only way to meet customer expectations for more personalized experiences.


Identity resolution matches a user’s multiple identifiers to build a complete and addressable customer profile across their product journey. This enables businesses to deliver relevant messaging and enhance the customer’s experience with their brand. The complexity of the device, platform and channel landscape means that brands need to be able to create a consistent brand experience for their customers – no matter what they are engaging with. Activating data through an identity resolution platform is critical to building this seamless, friction-free, optimal experience. Activation is also important to marketers who want to adhere to prevailing data privacy regulations such as GDPR and CCPA. These regulations define rules that must be followed when identifying, protecting and sharing personal information about a customer or household. For example, GDPR defines ‘personal information’ as anything associated with an individual or household, including device and app usage data, IP addresses and cookies. CCPA raises the stakes even more, as it empowers consumers to make a Subject Access Request that can expose a brand’s PII. Neural networks use backpropagation to adjust their weights and biases to minimize the cost function. This requires an activation function, a non-linear process that helps the neuron output become more accurate by adding extra non-linearity to the network. This activation function is typically based on properties from dropout, zone out and ReLUs (regularizers that stochastically multiply inputs by zero or one), yielding a neuron’s output value.


Identity resolution enables organizations to build a holistic view of customer identities across devices and touchpoints. This view provides a complete picture of consumers’ digital interactions that can be used to deliver relevant messaging and engagements. For example, if someone signs up for an account on a website, they may have been browsing the site as an anonymous user before creating an account. This information can then be combined with customers’ browsing data to give them more accurate recommendations. To accomplish this, an organization must first use a process to find and evaluate the quality of candidate-identifying attributes that enable an entity to be recognized as a distinct individual in different source systems. This process is often called entity reference extraction or named entity recognition (ER). The resulting graph of a customer’s digital footprint can then be combined with a company’s existing customer database, CRM and other data sources to create an omnichannel identity profile. This unified view can then be used to drive highly targeted and personalized marketing campaigns. The result is a more accurate and relevant customer experience across every touchpoint.


Matching is connecting records that refer to the same real-world entity. This is a vital part of entity resolution, also called record linkage or record deduplication in the database and statistics community. The matching process can be very simple or very complex, depending on your data ecosystem and the goals of your Identity Resolution strategy. There are several approaches to match data ranging from deterministic to probabilistic. Finding a rule or algorithm that makes sense for your needs is critical. The selection should be based on a combination of factors, including how many records you want to match and how sophisticated a solution you need. Another important consideration is the volume of customer data you have access to. If your company has exclusive access to all the customer data relevant to your business, you may not need to use a third-party solution. However, if your organization has limited access to all the customer data you need, it may be better to partner with a third-party identity vendor. They can provide access to rich truth sets that connect offline identifiers like customer IDs with stable online identifiers like hashed email addresses. Moreover, it is crucial to look for a provider that will regularly refresh offline identifiers. This ensures that all the information you need to create an accurate Single-Customer View is available.

Data Sources

Identity resolution is linking data from multiple sources to a single person. It is used by businesses to establish a customer’s identity and is used in various industries, including financial services, retail and transportation. A significant challenge in business is the amount of customer data duplicated and scattered across many data sources. For example, one person might have five different name variations, email addresses, phone numbers and physical addresses within their organization or on third-party platforms connected to them. Identifying a single person is critical to creating more personalized and relevant experiences across all customer engagements, from marketing to sales and loyalty. However, the challenge is that businesses often need help to merge all information into a single view. Fortunately, a first-party data strategy and identity graph solutions enable marketers to create a holistic customer profile with both stored and derived attributes, providing the best possible view of an individual in any context. This approach allows for better data quality and a higher personalization accuracy, reducing waste and expanding outbound reach to increase ROI for all stakeholders. In addition, it strengthens privacy Governance, Risk and Compliance (GRC) and supports consumer trust in brands and their policies and practices. It also improves closed-loop measurement, attribution and multi-channel campaign tracking.