Home Blog User Identification ...

User Identification Techniques: beyond Third-Party Cookies

Riccardo Sozzi

In the dynamic digital ecosystem, user identification is the cornerstone on which targeted and personalized marketing strategies are built. The ability to track, understand and engage users across various touch points is crucial to the success of any digital campaign.

Historically, cookies have been the primary tool for tracking user activity on the Web. Despite their usefulness, cookies have significant limitations: they are susceptible to deletion, have limited validity and, most importantly, do not work effectively in a cross-device or cross-platform environment.

With the introduction of stringent regulations such as GDPR in Europe and CCPA in California, the pressure for a more privacy-friendly and more reliable approach to user identification has intensified. In this context, new paradigms and technologies, including Persistent User ID, Probabilistic Fingerprinting, and Identity Resolution, have emerged that promise to overcome the limitations of cookies while ensuring compliance with privacy regulations.

Persistent User ID refers to a unique, stable, and persistent identifier that accompanies the user across multiple sessions and devices. Unlike cookies, which can be easily deleted and are limited to a single browser, Persistent User ID offers a holistic and long-lasting view of the user, making it particularly valuable for long-term marketing strategies. 

Probabilistic Fingerprinting, on the other hand, is a sophisticated approach to infer user identity by analyzing a set of device attributes and browsing behaviour. This method is less intrusive than cookies and considerably more resistant to identity erasure or masking measures.

Lastly Identity Resolution, the set of technologies and procedures that enable the reconciliation and unification of different identities or data fragments into a single, complete user profile. This process is critical in an era when users interact with brands across multiple channels and devices. Through sophisticated data matching techniques and machine learning algorithms, Identity Resolution enables companies to build a complete picture of the user, enabling highly personalized and measured campaigns. For example, a company could use Identity Resolution to link a user’s interaction with an email campaign with their in-app purchase behaviour, enabling unprecedented personalization.

In summary, as the digital landscape continues to evolve, user identification strategies are also adapting, shifting more and more toward solutions that ensure accuracy, persistence, and respect for privacy. Persistent User ID, Probabilistic Fingerprinting and Identity Resolution represent the frontier of this evolution, offering companies the levers to understand, engage and retain users in a way that was unimaginable in the cookie era.

Limitations of Traditional Cookie-Based Methods.

Cookies, though they have long been the mainstay of user identification and personalization online, have a number of inherent limitations that undermine their effectiveness in today’s digital landscape. These limitations not only hinder the ability of companies to accurately track and engage users, but also raise significant privacy and security concerns.

We can divide the limitations of cookies into two broad categories.

Technical Limitations:

  • Limited Lifespan: Cookies have a limited lifespan. Users can delete them manually, and modern browsers offer increasingly aggressive ways to limit their longevity.
  • Cross-Domain and Cross-Device Issues: Cookies work well within a single domain or website, but cannot track user activity across different domains or devices, thus limiting the user’s overall view.
  • Privacy Issues: Cookies are often perceived as invasive because they collect data on users without their explicit consent or understanding, raising concerns about privacy and data security.

Regulatory Limitations:

With the introduction of the General Data Protection Regulation (GDPR) in Europe and other similar regulations in various jurisdictions, the use of cookies has become significantly more complex. These regulations require companies to obtain users’ explicit consent before they can track their online activity with cookies, further limiting their usefulness.

Heavy penalties for noncompliance have prompted companies to look for alternative methods of identifying users that not only comply with privacy regulations but are also more effective and secure.

As these limitations have emerged, the market has begun to increasingly demand more resilient and privacy-compliant solutions.

Companies are demanding more transparent, secure and consent-based tracking methods. They are looking for user identification solutions that are not only compliant with regulations, but also provide a more accurate and unified view of the user across various touch points and devices.

This demand has driven innovation in technologies such as Persistent User ID, Probabilistic Fingerprinting, and Identity Resolution, which promise to overcome the limitations of cookies.

Such solutions offer new opportunities for more sophisticated and personalized user engagement and represent a step toward a more secure and transparent digital future.

Persistent User ID

Let us now focus on the technologies and approaches mentioned above, starting with persistent User-ID, a user identification technique that represents a real paradigm shift.

The advantages of using this type of technology are many. First and foremost is the ability to have very accurate profiling and targeting, thanks to a consistent view of the user across different devices and platforms.

Because Persistent User ID assignment often occurs with the user’s consent (e.g., during login), this method is generally more compliant with privacy regulations than traditional cookies.

The implementation of Persistent User IDs requires an advanced ID choice strategy and a robust identity management platform that can collect, store, and process identifiers in a secure and compliant manner.

We must start with the choice of the element that will serve as the persistent ID, which has characteristics of both recognizability and uniqueness while protecting the user’s privacy.

One of the most widely used techniques is the use of email hashed with the SHA256 algorithm. Indeed, hashing allows great security in protecting privacy while at the same time offering many possibilities for synchronization with both internal and external platforms.

We need to be able to quickly, securely, and efficiently generate the encrypted email and better manage exposure across all user touchpoints in order to consistently collect the data.

The platform chosen must be robust and flexible, and the setup strategy must be able to effectively expose the identifier across CRM, DataLayer, and Customer Data Platform. Exposure is often the most critical time of the persistent UserID, and its absence makes the whole setup of the 360-degree view of the user futile.

Probabilistic Fingerprinting

Probabilistic Fingerprinting emerges as a sophisticated and innovative technique for user identification. This method is distinguished by its ability to infer identity by analyzing a combination of device or browser attributes to create a unique “fingerprint.

Probabilistic Fingerprinting is distinguished by its probability-based approach. Instead of relying on a single identifier, as is the case with Persistent User ID, Probabilistic Fingerprinting analyzes a set of attributes to calculate the probability that a given browsing activity belongs to a specific user.

The main differences between a deterministic and a probabilistic approach are as follows:

  • Deterministic Fingerprinting relies on clear, unique identifiers, such as email address or phone number, to track users. Its accuracy is high, but it requires explicit personal data, raising privacy concerns;
  • Probabilistic Fingerprinting, on the other hand, relies on the analysis of behavioural patterns and device attributes, thus avoiding the direct collection of personal data. Although this approach is less accurate than the deterministic method, it offers a better balance between personalization and privacy compliance.

Probabilistic Fingerprinting uses device attributes such as hardware information the screen resolution, CPU and memory along with browser attributes such as user agent, language settings, installed plugins and cookie settings, Network attributes such as IP Addresses, HTTP headers and timezone settings and finally User behavior such as Typing Patterns, mouse movements and scroll speed.

There are some objective advantages to using probabilistic fingerprinting: 

  • Privacy Compliance: by reducing the need to collect direct personal data, it better aligns with privacy regulations;
  • Resistance to Deletion: unlike cookies, fingerprints cannot be easily deleted or blocked;
  • Cross-Device Tracking: ability to track users across different devices and browsers without the need for logins or persistent identifiers.

There is no shortage of challenges in this approach, which precisely because of its probabilistic nature implies a certain level of error. 

We can mention:

  • Variable Accuracy: the accuracy of Probabilistic Fingerprinting can vary depending on the quantity and quality of the attributes analyzed;
  • Technical Complexity: implementing a robust Probabilistic Fingerprinting system requires advanced technical skills and significant resources.

Retrospective User ID

Retrospective User ID reconciles previously anonymously collected data with a unique identifier as soon as the user provides personal data and consents to the tracking. 

This method represents a bridge between anonymity and personalized identification, preserving the user’s privacy until the moment the user decides to reveal his or her identity.

In detail, when a user interacts with an online service without providing personal data, his or her activities are tracked and stored anonymously, associating them with a unique probabilistic fingerprint based on various attributes such as device settings, browsing behaviour, and other digital signals. This process ensures a basic level of personalization and behavioural analysis without compromising the user’s identity.

The true potential of Retrospective User ID manifests itself the moment the user decides to share personal information, such as an email address or phone number, along with consent for tracking. At this instant, all previously collected data associated with the probabilistic fingerprint can be retrospectively linked to the user’s identity. This creates a unified view of the user’s path, enriching data analysis and providing much more targeted and effective personalization opportunities.

This technology not only respects user privacy, but also offers significant benefits to businesses. It allows them to maintain an in-depth understanding of user behaviour without requiring personal data from the outset, enabling them to build a relationship based on trust and conscious acceptance of tracking by the user. In addition, Retrospective User ID opens new avenues for sophisticated marketing strategies, such as advanced personalization and predictive analytics, based on a complete and accurate picture of user behavior throughout the user’s digital journey.

Identity Integration and Synchronization

All of the above technological elements are intended to enable effective user reconciliation. Identity resolution is when we recognize the user in our systems and begin the process of actually activating the user’s business. This methodology aims to unify and make the different user identities consistent, transforming fragments of data into complete and accurate user profiles.

In the context of digital marketing, Identity Resolution is critical to ensure that targeting and personalization strategies are based on a complete and accurate understanding of the user.

It provides the foundation for advanced personalization, enabling marketers to deliver relevant and consistent messages across all user touch points.

Accuracy in Identity Resolution is crucial to ensure that communications and offers are relevant. Errors or inaccuracies can lead to poor targeting, damaging brand reputation and campaign effectiveness.

Once the user’s identity is resolved through fingerprints or unique identifiers, we can integrate online and offline events by merging data from interactions on websites or mobile apps with those originating from interactions in physical stores, call centers, and so on.

Through identity resolution, then, we can use data matching techniques to combine, compare, and unify data from different sources.

We have several strategies to resolve user identity. We can rely on automation and Machine Learning by using algorithms that can automate data integration and cleansing, improving accuracy and reducing effort.

Downstream of a user-focused strategy, we can really appreciate the value of customer-centric solutions such as Customer Data Platforms.

These technology solutions are gaining momentum in companies because they enable advanced strategies that have been put under severe strain by the end of third-party cookies and new privacy restrictions.

Today, an agile, headless approach is preferred, which instead of involving the complete installation of a CDP, builds the Customer Data Platform on top of the present cloud-based enterprise database, avoiding data duplication, inefficiency in identity resolution, and maintenance complexity.

Within an agile, composable platform, we can choose the User-ID, Fingerprinting and Retrospective User ID techniques we prefer, instead of being forced to use the native systems of the integrated solution, which are often excellent on one point and potentially deficient on others.

User ID is a key building block for effective data analysis in an Analytics platform. By tracking users across different devices and sessions, it allows us to gain a holistic view of their behaviour and activate personalized and targeted marketing strategies.

By configuring an advanced user data recognition and reconciliation system, we can enable several use cases:

  • Personalization of User Experience: E-commerce platforms can leverage user information to offer product recommendations based on previous interactions and preferences, regardless of the device used, thanks to the data collected or prediction through artificial intelligence.
  • Targeted Advertising: Advertisers can leverage ID to build advertising campaigns that reach the same user on different devices, thereby increasing the consistency and effectiveness of communication. At the same time, advanced segmentation makes it possible to create statistically representative segments of the best customers for Audience Seeding strategies of Look-Alike algorithms, overcoming the increasingly obvious problem of difficulty in targeting new customers.
  • User Behavior Analysis: Combining data from different sessions and devices provides more granular and reliable insights into users. Analytics platforms track their journey across multiple touchpoints, providing a holistic view of the customer journey. The use of custom dimensions enriches the analysis with extra information, allowing you to:
    • Understand the customer journey: identify strengths and weaknesses to optimize the user experience and increase conversions.
    • Personalize the user experience: create targeted marketing campaigns and offers, increasing relevance and engagement.
    • Identify new business opportunities: detect new trends and potentially interesting market segments.