In the world of digital marketing and data analytics, two key methods used to understand the effectiveness of marketing efforts are Attribution Modeling and Marketing Mix Modeling (MMM). As a beginner data analyst, you may come across these terms frequently, but understanding the nuances and applications of each can be challenging. In this blog post, we’ll break down both concepts, highlight their differences, and explain when and how they are used. What is Attribution Modeling? Attribution modeling is a method used in digital marketing to assign credit to different marketing touchpoints that a consumer interacts with on their journey toward a conversion. In simpler terms, it helps marketers understand which channels, ads, or campaigns should be credited for driving a sale or a lead. Types of Attribution Models: There are various types of attribution models, each with a unique way of distributing credit. The most widespread are the following: First-Touch At...
The Problem of Shared Devices in Identity Resolution for Customer Data Platforms
Understanding customer behavior is paramount for businesses striving to provide personalized experiences in the digital era. Customer Data Platforms (CDPs) play a critical role in this by aggregating data from various sources to create comprehensive customer profiles. However, one significant challenge that arises in this process is identity resolution, especially when dealing with shared devices. Unfortunately, many businesses recognise this problem quite late during the implementation problem, or even worse just after the go live. It can be then far more costly to address it, so better be aware about the cause, implications and possible solutions right at the start. That is why this blog post tries to give the first insight into the problem connected to common identity resolution process.
The Complexity of Identity Resolution for Customer Data Platforms (CDPs)
Customer Data Platforms are sophisticated systems that collect, consolidate, and unify customer data from multiple sources such as internal databases, websites, mobile apps, social media, and offline interactions. The goal of a CDP is to create a 360-degree view of the customer, enabling businesses to activate advanced audiences and to deliver personalized and targeted marketing activities to them. Effective identity resolution is crucial for achieving this goal, as it ensures that all data points collected to a single individual are accurately attributed and connected to customer profile.
Identity resolution is the process of matching and merging data points from various sources to a single customer profile. This involves reconciling different identifiers such as email addresses, phone numbers, CRM IDs, device IDs, and values of cookies. The challenge becomes more pronounced when multiple individuals use the same device, as is common in households, public computers or shared workspaces.
The Problem of Shared Devices
Shared devices introduce significant ambiguity into the identity resolution process. When multiple people use the same device, distinguishing between their interactions and behaviors becomes difficult. For instance, consider a family sharing a tablet. The browsing history, app usage, and other online activities on that device could belong to any family member, making it challenging for a CDP to accurately attribute actions to the correct individual. Even if the users are logged in, i.e. some kind of CRM ID is present in the collected data, due to the out of the box deterministic identity resolution algorithms typical for many CDPs, the profiles will be stitched together because of the presence of the same cookie based ID or device ID.
Impact on Customer Profiles
The ambiguity caused by shared devices can lead to inaccurate customer profiles. A CDP might incorrectly merge data from different users into a single profile, resulting in a distorted view of customer preferences and behaviors. This, in turn, can affect the effectiveness of personalized marketing efforts, leading to irrelevant recommendations and potentially eroding customer trust. Moreover, shared devices also raise privacy concerns. When a CDP cannot accurately differentiate between users, it might inadvertently expose sensitive information to unintended recipients. For example, personalized ads or recommendations meant for one individual might be visible to other users of the same device, compromising trust and privacy.
Strategies for Addressing Shared Device Challenges
Addressing the challenges posed by shared devices requires a multi-faceted approach that combines technological advancements, data management practices, and user education. Here are some strategies to consider:
1. Contextual Signals
Leveraging contextual signals can help distinguish between users on a shared device. These signals include location data, time of day, and usage patterns. For example, if a device is typically used by one person during working hours and another in the evenings, these patterns can provide clues to differentiate between users.
2. Behavioral Analytics for Probabilistic Merging Policies
Advanced behavioral analytics can play a crucial role in identity resolution. By analyzing user behavior patterns such as browsing habits, app usage, and interaction frequencies, CDPs can develop more accurate profiles. Machine learning algorithms can identify subtle differences in behavior that may indicate different users on the same device.
3. More Advanced Conditional Deterministic Merge Policies
Some CDP vendors who are aware of the problem try to implement new options how to adjust their deterministic merge policies, so additional conditions can be applied to better leverage the presence of person based identifiers, e.g. CRM IDs. The knowledge from already created profiles can be used back in identity stitching to prevent the unwanted profile merging.
4. Management of Cookie Based and Device IDs
Another option is more sophisticated management and set up of cookie based or device based IDs. As these values are cause of profiles stitching on same device, the more control about how and when the values are passed into the CDPs can improve the accuracy for identity stitching. However, this approach can then have another consequences, e.g. on another tools reliant on their values. On the other hand, if done in the proper way, also these additional tools can leverage the profiling of CDP back and can improve the accuracy of their own user recognition process.
Conclusion
The problem of shared devices presents a significant challenge for identity resolution in Customer Data Platforms, even though it can be hidden at the first sight. The ambiguity introduced by multiple users on the same device can lead to inaccurate customer profiles and privacy concerns. However, there are ways how businesses can mitigate these challenges and improve the accuracy of their CDPs. Therefore it is crucial to access during the testing phase of implementation projects, how big issue this problem is for particular CDP customer. Then, the proper solution can be selected and tested. It is also crucial for businesses to stay informed about the latest advancements and best practices in this field to ensure they are delivering the best possible customer experiences while maintaining trust and privacy. By doing so, they can harness the full potential of their Customer Data Platforms and achieve their marketing and business goals.