Identity: Evaluating Tech Categories
I recently suggested a framework to help think about identity in the context of customer experience (CX).
However, evaluating technology for identity is complex. What category of software should you consider when evaluating identity tech?
The answer is nuanced.
We will examine four categories of technology to consider for identity initiatives:
Master Data Management (MDM)
Customer Data Platform (CDP)
Identity Provider
Data Warehouse
Master Data Management (MDM)
Master Data Management (MDM) solutions have been around for a long time. They are focusing on identity across the enterprise, beyond just the customer data domain. MDMs provide tools that allow IT to ensure the uniformity, accuracy, stewardship, and accountability of the enterprise's master data assets.
MDM in a nutshell
Users: Resolve identity for authenticated users, using PII.
Channels: Primarily used for owned channels (email, direct mail). However, I predict that paid media channels will increasingly benefit from the brands’ work to resolve identities with MDMs.
ID Source: Exclusively working with a brand’s 1st-party data.
Technique: Use a combination of deterministic and probabilistic techniques.
Latency: While most organizations focus on batch identity resolution with their MDM, many vendors have added real-time capabilities to their offering.
Recommendations
MDMs are best for organizations with the people and budgets to deploy and maintain advanced identity resolution or those with strict regulatory obligations around identity (e.g. healthcare, finance). MDMs are often used for the creation of a trusted “golden record.”
While MDMs can be expensive to deploy and maintain, they provide the greatest level of flexibility and self-service controls for enterprises looking to manage identity around their 1st-party data.
The results of this identity management often integrate into systems beyond customer experience (CX), such as billing systems, invoicing, compliance, etc.
Customer Data Platform (CDP)
Customer Data Platform (CDP) solutions were originally created to help business teams to become self-serving and customer-centric with their CX use cases, by unifying all customer data and providing a marketer-friendly interface.
CDP in a nutshell
Users: With Data Management Platforms (DMP) going away, CDPs are increasingly used for anonymous users, although their focus has historically leaned towards authenticated users.
Channels: Often integrate with both owned channels and paid media.
ID Source: Identity resolution is applied to 1st-party data sets.
Technique: While every CDP will propose some deterministic identity resolution, only a few add probabilistic techniques. These usually offer less self-service controls than the MDM alternative.
Latency: Options vary widely across the CDP market, batch resolution being more prominent than real-time.
Recommendations
A CDP is best for organizations with comprehensive identity resolution needs who are primarily concerned with enhancing the customer experience of their marketing activities.
When an organization lacks the team, budget, or skills to deploy and maintain an MDM, a CDP identity resolution product is a good alternative to consider. However, be aware that capabilities around identity resolution vary greatly among CDPs.
Identity Provider
Identity Providers enable the connection, enrichment, and control of identity data by linking first-, second-, and third-party data. They combine addressable media signals with consumer attributes and behavior, typically resulting in a large third-party identity graph accessible to brands.
Identity Provider in a nutshell
Users: Onboard and make both anonymous and authenticated users addressable.
Channels: Exclusively focused on addressability for paid media channels
ID Source: Identity resolution is applied to 1st-party data sets.
Technique: Deterministic matching is standard, with some lightweight probabilistic techniques can be applied.
Latency: Some vendors may be able to stitch records in real-time, but most identity resolution is achieved in batch, with a low update frequency.
Recommendations
An Identity Provider is best for improving CX when there is limited access to 1st-party data and/or when looking to increase addressability. As anonymous identities become more challenging to access and tie together, brands will likely see a drop in the quality and efficiency of identity providers for non-authenticated visitors.
Ensure that the identity provided by the vendor offers addressability and value for your specific use cases and channels.
Data Warehouse
Up until recently, considering the Data Warehouse as an option for identity use cases meant having a database engineer or data scientist custom-build a solution from the ground up. However, cloud data warehouses have recently started announcing their identity services to help accelerate this work.
Data Warehouse in a nutshell
Users: Similar to CDPs, mainly for authenticated users but can also be used for anonymous users.
Channels: The natural channels for data warehouses are owned channels, but they are increasingly looking to play a role with paid media.
ID Source: Customers are connecting their 1st-party against the identity graph from the provider, which is almost exclusively formed using 3rd-party data.
Technique: Each vendor has different “preferences”, and the technique used might differ for their anonymous vs authenticated ID graph. Probabilistic matching in this scenario indicates the use of fingerprinting techniques to identify anonymous users.
Latency: Exclusively focused on batch processing.
Recommendations
A Data Warehouse is best for organizations already heavily investing in centralizing their data in a cloud data warehouse and looking to build or integrate identity solutions in-house. Due to the recency of identity services from data warehouses, significant effort may be required to use the warehouse for identity resolution, unless an application sitting directly on top of it is selected.
Connecting the solutions to your stack
🔑 Before settling on any technology, make sure to understand the current pain points you are looking to address.
💡 As you explore more options, you'll realize that no single piece of technology will satisfy all your needs. In fact, it’s not uncommon for a brand to have all four options we discussed today within their stack.
While MDMs and Identity Providers are solely focused on identity use cases, CDPs and Data Warehouses are often used for a broader scope.
Finally, within each category, vendor capabilities and approaches vary greatly.