Over the past two decades, digital analytics practitioners have perfected the science of measuring and analyzing digital experiences. Best practices and standardized metrics have been established by industry groups. Some of those metrics have even achieved the level of organizational KPIs. While the emergence of mobile offered a brief speed bump, the industry consolidated around a generally accepted way of doing digital analytics.
At the same time, there was a sense that traditional digital analytics didn’t always provide a complete picture of a customer’s journey across channels beyond web and mobile apps. Additionally, an increased focus on consumer privacy has gradually chipped away at the completeness of the digital experience data that can be collected. This has required data teams to increasingly look inward and rely on first-party datasets that exist within an organization.
Some intrepid teams set out to expand horizons by exporting digital analytics data feeds into Hadoop or the cloud to combine it with other customer data from CRMs, call centers, physical stores, etc. Ultimately, the sheer size of digital data, its fast-changing nature, and unpredictable data profile made it difficult to model and even harder to explore using standard SQL and BI tools.
Enter Adobe’s latest version of its Analytics product: Adobe Customer Journey Analytics (CJA). Adobe has better equipped CJA to handle and represent customer data that extends beyond digital while retaining the no-code Analysis Workspace UI that made it so easy to quickly explore and uncover insights without any knowledge of SQL. That said, CJA’s expanded scope comes with some big changes under the hood with the potential to impact stakeholders across the organization. We’re going to discuss the potential pitfalls you may come across if you’re considering or undertaking a CJA migration, as well as some suggestions for how to navigate them.
At the heart of CJA’s new architecture is Adobe’s data lake contained within Adobe Experience Platform (AEP). CJA consumes all of the data it uses from the AEP data lake. To move data in and out of AEP, Adobe created its own open-source data model named Experience Data Model, or XDM (so many acronyms).
In comparison to the flexibility of props and eVars, it may seem like XDM schema for CJA data structures is overly strict. On the other hand, the interaction with more rigorously modeled and potentially sensitive customer data from other data lakes and warehouses has made schema a necessary prerequisite to meaningfully and safely onboard external data. For all of these disparate datasets to blend seamlessly in a reporting environment, there have to be consistent and clear definitions across the datasets.
This level of data modeling can be a shock to the system for Adobe Analytics users, where it was common practice to repurpose eVars and change their semantics — an approach that would be strongly discouraged in traditional data modeling. As digital analytics data is elevated into a critical organizational asset to measure and execute customer journeys, schema can help facilitate the level of data quality and fidelity necessary to fulfill those use cases. A side effect of this is a need for greater emphasis on validation and accuracy in data collection. If an event collected from the client doesn’t conform to the expected XDM schema, it will fail upon ingestion into AEP.
Lastly, a benefit of CJA supporting schema-based data structures is not being locked into one vendor-determined schema, like most digital analytics tools. This means data collected from other digital analytics tools, including GA4, can easily be ingested and reported on in CJA.
Key Takeaway: Fully representing the customer journey will require having believable data and XDM schema goes a long way toward ensuring trust. As you set off to implement CJA, chatting with resident data architects and engineers to pick their brains on data modeling or seeking some of their services will be time well spent.
Moving past AEP, a successful CJA implementation starts and ends with identification — the backbone of a customer journey. When setting up CJA, every event dataset that is brought in will require a common Person ID. The selection of a Person ID is possibly the single most impactful decision you’ll make within your CJA implementation.
As you evaluate which data sources to bring into CJA, your organization likely has many different identifiers across its various systems. Picking one as your Person ID may mean leaving some data sources out of your CJA implementation until system owners can add the preferred Person ID.
For web and mobile data, if you have an authenticated experience, you may also need an additional plan to smooth the changing levels of identification across your experience. Identity stitching is an effective way to ensure that sessions aren’t broken when a new identifier suddenly becomes available upon authentication.
Key Takeaway: The Person ID you choose can make or break your CJA implementation. Take the time to evaluate the sources you want to include and determine an ID strategy that will best fit your implementation.
With almost limitless possibilities, deciding what data to bring into CJA can quickly become overwhelming. If your initial deployment is going to be successful, you’ll need to solidify what minimum viable product (MVP) you can roll out to users. Follow these steps to help define what that is:
As you go through these steps, it’s likely that you’ll find that certain data sources you wanted to bring in from your wish list aren’t feasible in the near term. Put those in the backlog and instead focus on what you can deliver as an MVP. That may mean that the first iteration of your CJA implementation is only digital analytics data plus one or two other data sources and that’s okay. You can always add more data sources in future iterations.
Key Takeaway: To get CJA off the ground, align on an achievable MVP and don't settle for less than end-to-end integration. Manual or ad hoc uploads will haunt your implementation as tech debt that will undermine the trust of your users.
When moving from digital to journey analytics, the impact on KPIs can be one of the more jarring outcomes, particularly Visitors and Visits. So different, in fact, that Adobe opted to rename the default metrics in CJA to People and Sessions to avoid equating the two sets of metrics.
Depending on the type of business, level of identification and data sources being brought in, the differences between metrics in Adobe Analytics and CJA can be substantial. This shouldn’t be a complete surprise as the tool measures a different entity: customer vs. browser/device. Also, the shifts will not consistently be up nor down. For example, if you have a lot of authenticated users moving between a mobile app and mobile web, you will see those sessions collapse down. Meanwhile, if you are bringing in standalone event data like appointments, you will likely see an increase in sessions.
Lastly, don’t forget that calculated metrics derived from Visits and Visitors, such as Conversion Rate, will also undergo significant shifts. First impressions are everything for a major product migration. If you don’t properly set expectations, these shifts can quickly send executives into a panic, and it will take twice as long to build back their trust.
Key Takeaway: Prepare for an entirely new baseline for KPIs. Monitor how they differ from the previous metrics and develop a communication plan to proactively answer questions and help stakeholders understand why there will be shifts.
Implementing CJA is no small task. It will involve significant time and planning. You don’t want to put forward an aggressive timeline with endless delays and harm your reputation before launch. For larger organizations, a year or more is a reasonable timeline to fully wind down your legacy digital analytics solution.
If you’re not already practicing agile development, adopt some key principles used by successful agile data products: stay focused on the MVP, break tasks down into smaller, more manageable pieces and provide regular updates and demos to stakeholders.
Key Takeaway: Provide realistic timelines and over-communicate. The more engaged your users are, the more understanding they’ll be if timelines slip.
With how similar Adobe CJA looks and feels to Adobe Analytics, it would be easy to get complacent and bypass a training plan for your migration. On the contrary, you shouldn’t underestimate the under-the-hood changes that will cause users to experience the data in a much different way.
For existing analytics users, tailor your training to focus on the data instead of the basics of how to use the tool. Identify a group of early adopters that can participate in an early access program and regular office hours. Gather their feedback and incorporate it into future iterations. These users will be critical to flush out issues, develop new best practices and advocate for the product.
On the flip side, new data sources likely means new users that would find value in becoming familiar with Adobe CJA. Interview any and all analytics professionals who may benefit. These users will likely be coming from a SQL or BI background, so ensure they get a more detailed training on how to take advantage of working in Analysis Workspace. Even more important, listen to them and be sensitive to concerns they may have about discrepancies reported from the systems they’re familiar with. Fostering a community of contributors will make for a healthier implementation and build trust within the organization.
Key Takeaway: The migration to Adobe CJA is a golden opportunity to maximize value for your organization by building loyalty among existing users and expanding the number of new users in the tool—don’t pass it up!
Despite on-the-surface similarities in Analysis Workspace, a move from measuring digital experiences with Adobe Analytics to customer journeys with CJA will have dramatic impacts that can ripple across your organization. It won’t be a simple code change and upgrade. Measuring the customer’s full journey requires a major organizational shift in mindset. A lot of the rules that applied for digital analytics may not apply in a journey context or will need updating. The combination of data sources your organization brings into CJA may be entirely unique, so expect uncharted waters and encourage innovation and collaboration among users.
With the changing privacy landscape, there is an opening for organizations that can tear down internal silos and exploit the first-party data that they already have. Time to insight, data democratization and self-service have become buzzwords that can actually be delivered with a quality Adobe CJA implementation. With the proper planning and balance of change management, the migration from digital analytics to customer journey analytics can be the accelerator your organization is looking for.
Do you have more questions or would like to learn more? Schedule a 30-minute consultation to discuss the details of your Adobe CJA migration.