What is identity resolution and how are platforms adapting to privacy changes?

Identity resolution  — the science of connecting the growing volume of consumer identifiers to one individual as they interact across channels and devices – has become critical to marketing success and essential for compliance with consumer privacy laws such as the California Consumer Privacy Act (CCPA) and the European Union’s General Data Protection Regulation (GDPR).

Identity resolution platforms sit at the heart of this discussion. These software tools connect consumer identifiers across different channels and devices. They do this accurately, at scale, and while respecting privacy. This creates a lasting and addressable individual profile. Identity resolution platforms help marketers “close the loop” in customer marketing and provide a complete view of customer activity across all touchpoints and channels in an organization. These identifiers can and should encompass both online (device, email, cookie or mobile ad ID) and offline (name, address, phone number) data signals and attributes.

In this article you’ll learn:

• Why do marketers need identity resolution platforms?
• What identity resolution platforms do
• What is data onboarding and why is it important?
• What is an identity graph?
• Why privacy compliance and data ownership matters
• What third-party software integrations are available for identity resolution?
• What are the benefits of using identity resolution platforms?
• FAQs and further reading

Why do marketers need identity resolution platforms?

The number and variety of devices that can connect to the internet in 2024 is impressive. Consumers can interact and usually purchase from brands using a computer, a mobile phone, a gaming console, a Smart TV, a VR headset or a home assistant like Google Nest or Amazon Alexa. In fact, there are smart refrigerators that can order groceries according to your preferences, when you’re running low. 

In this competitive environment, it is essential that brand marketers understand which online devices and offline behaviors belong to a consumer as well as who that consumer is. Every time a consumer interacts with the brand — regardless of channel – a different identifier (also called a key) can be attributed to that individual. These identifiers can include an email, IP or physical address, as well as a mobile phone number, digital tag or cookie.

But accurately resolving consumer identities has proved challenging for a majority of brand marketers. Forrester found that a majority of C-level executives overrate their marketing organization’s customer identity accuracy and persistence.

That challenge promises to become even more difficult as tech companies make changes that essentially deprecate third-party cookies — one of the key identifiers used to stitch identity data together. Google began the phased roll-out of cookie deprecation earlier this year, before deciding that the choice regarding third-party cookies would be up to the consumers. Apple has similar plans for IDFA, its identifier for advertisers, requiring users to opt into the program. Additionally, since the roll-out of iOS 17 in late 2023 the new Link Tracking Protection feature poses further challenges for marketers. 

Identity resolution is not only critical to marketing success but is essential for compliance with consumer privacy laws such as CCPA and GDPR. Explore the platforms essential to identity resolution in the latest edition of this MarTech Intelligence Report.

Click here to download!

What identity resolution platforms do

Identity resolution platforms support marketing processes around targeting, measurement and personalization for both known and anonymous audiences across digital and offline channels. And most enterprise identity resolution platform vendors offer the following core features and capabilities:

• Data onboarding (including online/offline matching).
• Proprietary identity graph.
• Client ownership of first-party data.
• Persistent individual and/or household ID.
• Compliance with privacy regulations.
• APIs for third-party system integration.

Vendors begin to differentiate their platforms by offering more advanced features, sometimes requiring additional investment, which include – but are not limited to – the following:

• Match confidence scoring.
• Data clean rooms.
• Private (first-party) and/or second-party cooperative identity graphs.
• Pre-built connections to martech/ad tech platforms.

Let’s look deeper at these platform capabilities..

What is data onboarding and why is it important?

Data onboarding is the first step in the identity resolution process. Client data is typically onboarded via secure file transfer (SFTP), although several vendors also provide direct API transfer or pixel syncs. Data is processed with the goal of establishing a universal view of the customer and includes the following:

• Matching individual identifiers in the identity graph (see below) to associate the customer with their interactions across touchpoints, particularly online to offline.
• Suppressing unresolved IDs and interaction data for potential future use.
• Hashing or tokenizing personally identifiable information (PII) with an anonymized customer ID.
• Linking matched IDs to a universal ID representing the customer profile and all of its associated attributes.
• Validating the accuracy of matches to a pre-established “truth set” of referential data known to be precise and accurate.

Most vendors provide persistent customer IDs during the identity resolution process, which means the ID follows the individual (or household) even as identifiers change, which they inevitably do. For example, when browser cookies expire, are deleted or customers buy and use new devices, the customer ID will remain the same. Persistence is also critical to enabling temporal time-series analytics, such as churn analytics. Matching algorithms differ among vendors, with matches established via probabilistic or deterministic methods or a combination of both. Deterministic matching relies on explicit links between identifiers, such as an email address that is used to sign in to a website or mobile app and can be associated with the resulting cookie or mobile ad ID (MAID). Probabilistic matching relies on implicit links between identifiers, such as a desktop cookie and MAID both associated with a residential IP address. The goal is to consider multiple signals like location and browsing history.

Both approaches have their pros and cons, which should be considered when choosing an identity resolution platform. Deterministic matching takes an omnichannel view that attempts to connect identifiers across digital and offline interactions. It can be difficult to scale and prone to inaccuracy. Probabilistic matching can “weed out” inaccurate data because it looks at a variety of data points versus binary matches. Its drawback is that it is limited to online touchpoints. Some vendors are using hybrid identity resolution approaches, which try to compensate for deterministic and probabilistic weaknesses while capitalizing on their advantages. It uses deterministic and probabilistic linkages, and then merges the two linkage sets together to form new, combined clusters.

Many vendors provide their overall match rates to potential clients. A few vendors go a step further, providing clients with customizable match algorithms or confidence scores (how likely the matches are accurate) based on their specific first-party customer data and data quality profiles. For example, a pure online organization may rarely use postal addresses and is likely to have lower-quality address data than an organization that relies on fulfillment to a physical shipping address. Addressability is another factor that can help marketers measure their match accuracy by assessing the number of consumers that can actually be contacted.es and is likely to have lower-quality address data than an organization that relies on fulfillment to a physical shipping address. Addressability is another factor that can help marketers measure their match accuracy by assessing the number of consumers that can actually be contacted.

What is an identity graph?

Most identity resolution vendors maintain a proprietary identity graph or database that houses all the known identifiers that correlate with individual consumers. There is no standard model for an identity graph. Each vendor differs in the types of foundational PII used, the matching methods employed and the non-PII integrated to enrich the individual profiles. Across the buyer’s journey, many identifiers can be associated with an individual, including email addresses, physical addresses, landline and mobile phone numbers, mobile ad and device IDs, account usernames and loyalty numbers. The identity graph collects these identifiers and links them to customer profiles, which are used to target and personalize marketing messages.

Identity graphs may also incorporate demographic, behavioral, financial, lifestyle, purchase and other data compiled or licensed from third-party sources, such as online news sites, purchase transactions, surveys, email service providers (ESPs), motor vehicle records, voter registration and other public records. Having all of this customer device, channel and behavioral data in one place allows brand marketers to more accurately measure the reach and frequency of their campaigns, and analyze how different ads and marketing tactics perform across channels.

In response to the dwindling availability of third-party cookie data and the increasing use of consumer privacy tools, such as advertising and location blocking apps, several identity resolution platform vendors are offering identity graphs built on first-party or second-party datasets. First-party identity graphs are exclusively used by a brand to house and match known customer data. Second-party identity graphs use cooperative data-sharing agreements between multiple brands or publishers to create common, anonymized identity assets.

Participating organizations can build, plan, activate and measure custom audience pools to either target or suppress customers across addressable media.

Why privacy compliance and data ownership matters

Marketers with customers in the European Union have had to comply with GDPR since May 2018. The CCPA, impacting all brands with customers residing in California, went into effect in January 2020, and empowers consumers to make a Subject Access Request to see all the data an organization has about them, which raises the stakes of identity resolution match accuracy. CCPA defines personal information as anything that can be associated or linked with an individual or household.

Marketers in the highly regulated healthcare market must follow Health Insurance Portability and Accountability Act (HIPAA) and Health Information Technology for Economic and Clinical Health Act (HITECH) regulations. In addition, all organizations that accept, process, store or transmit credit card information must maintain a secure environment that meets Payment Card Industry Data Security Standards (PCI DSS), as well.

These regulations are driving an expanded industry focus on data transparency and consumer consent, with a view toward complying with new standards for the benefit of consumers, as well as marketers. Many identity resolution platform vendors adhere to advertising industry guidelines from the Digital Advertising Alliance (DAA) or Interactive Advertising Bureau (IAB).

Lastly, and importantly, the majority of vendors profiled generally allow enterprise brands to retain ownership of their first-party data.

What third-party software integrations are available for identity resolution?

The ultimate marketing goal for identity resolution is to support and enable data activation by pushing segmented audiences into highly personalized campaigns through a variety of martech (CRMs, DMPs, marketing automation platforms (MAPs), ESPs, etc.) and ad tech (DSPs, SSPs, ad exchanges, etc.) tools and platforms. Identity resolution platforms should be able to streamline integration with the client’s martech and ad tech ecosystems by providing pre-built (or native) connections and an extensive set of APIs for custom integrations. Access to these APIs may or may not be included in base pricing.

Explore platform capabilities from vendors like Acxiom, Experian, Infutor, Merkle and more in the full MarTech Intelligence Report on identity resolution platforms.

Click here to download!

How AI can assist in identity resolution

AI and machine learning (ML) are enhancing identity resolution platforms by improving the accuracy of matching customer data across multiple sources. ML algorithms can analyze large datasets to identify patterns, helping to match inconsistencies. These algorithms also clean, standardize and correct data errors while detecting anomalies. Predictive modeling helps forecast customer behavior, predicting whether different data points belong to the same person.

ML is also useful for processing unstructured data, like social media posts and emails, converting it into structured data that can be used for identity resolution. Natural language processing (NLP) extracts meaningful information from text, further improving identity matching.

Generative AI, particularly through generative adversarial networks (GANs), can address incomplete or sparse datasets by generating supplementary data that appears realistic. One part of the GAN creates new data while another evaluates its accuracy, leading to improved data generation over time.

These technological advancements bring additional benefits, including faster data processing, which allows marketers to react quickly to changes. Moreover, generative AI uses unified customer profiles from identity resolution platforms to craft personalized messages, enhancing marketing efforts by providing more relevant content. As AI and ML technologies continue to evolve, their ability to improve data accuracy, efficiency and personalization in identity resolution systems becomes increasingly valuable for marketers.

What are the benefits of using identity resolution platforms?

Connecting consumer identifiers has become a mandate for enterprise marketers trying to meet and exceed customer expectations for a consistent and personalized brand experience.

Automating the process with an identity resolution platform can provide the following benefits:

• Deeper customer insights. Piecing together data signals from multiple data sources and interactions enables marketers to build more robust customer profiles. Knowing the customer at a more granular level can help drive rich customer insights that enhance campaign targeting, personalization and relevance.
• Personalization accuracy. Better personalization is a primary marketing use case for many identity resolution platforms, which create a consistent set of identifiers to fuel personalized interactions. If you don’t know with confidence who your customer is, you can’t personalize your messages or experiences.
• More seamless customer experiences. Automated identity resolution allows marketing organizations to create a unified view of customers, which can be communicated and deployed across brands, business units and product lines. Recognizing customers across every step of the customer journey reduces waste by eliminating duplicate contacts, and enhances their experiences by enabling interactions in the right channel at the right time.
• Stronger privacy Governance, Risk and Compliance (GRC). Effective identity resolution supports your organization’s commitment to data governance and, ultimately, consumer trust in your brand. Using an identity resolution platform makes customer preference management (including opting out), as well as regulatory and corporate policy compliance easier and more comprehensive.
• Enhanced cross-channel attribution and campaign tracking. Persistent IDs that identify customers (both known and anonymous) across channels enables more accurate, closed-loop measurement and multi-touch attribution.
• Improved marketing ROI. Identity graphs reduce data overlap and duplication, resulting in more efficient spending on campaigns that work. Conversely, not knowing who your customers are leads to misidentifying them and engaging in ways they may perceive to be intrusive or irrelevant.

FAQs and further reading

I’m the person-in-charge of identity resolution at our organization. What are the accepted statistical models I can use to back-up data confidence with a framework? 

In general there are two widely standardized models of data accuracy, which are deterministic, vs. probabilistic matching. You may find that identity resolution software favors one of these models, though you can apply your own.

We’re a small organization with less than 1,000 customers. Do we need a formal identity resolution platform or process?

There’s a lot to be said for future-proofing in a climate of increasing privacy, even if your customer growth plans are modest. That said, there’s always a cost-benefit concern. Ask yourself these questions to assess if you’re ready for a software solution to resolving identities.

We’ve just invested in a CDP. Do I still need a separate identity resolution platform?

Generally speaking, most CDPs incorporate solid identity resolution capabilities. In fact identity resolution should underpin all aspects of all customer data systems. Here’s how our MarTechBot responded to a question about the difference between CDPs and identity resolution solutions.