Insights on Tealium Digital Velocity 2025

I was able to attend the Digital Velocity Tealium conference in London this week, and I thought it would have been nice to give you some insights on it. First things first:

• What is Tealium? Tealium is a platform for data acquisition and segmentation. In Sunrise B2C, we're using it mainly as a CDP.

• What is a CDP? A Customer Data Platform is a software in which we can collect user data from multiple sources and use automated or manual rules to segment this data. Various integrations (e.g. Google Firebase, Adobe Target) allow us to perform action on such segments of users (e.g. displaying a specific offer, sending a push notification, etc.)

Now, what was the conference about: 

• There was a lot of buzz regarding AI (especially GenAI), but I found limited applications of it. Some AI-powered use cases Tealium proposed us were:

  • Propensity Modeling: creating a real-time propensity data pipeline

  • Next Best Action: creating a real-time NBA data pipeline

  • Product/Service recommendations: creating a real-time recommendation data pipeline.

• Other GenAI use cases were focused on scaling content (check out the slide from Bahia Principe and Dynamic Yield)

• Particular focus was given to the Moments API. Through this API, we can access the data of a user easily, being able to customize their experience in real time across totally different domains (both for content personalization but also for some more innovative things, e.g. feeding this data in an AI-powered chatbot to give the user a tailored assistance) 

• This goes hand-in-hand with the evolution of processing data from batching to streaming: acquiring real time data and provide on-the-fly enrichments, all powered by APIs

• Tealium introduced a very sophisticated reporting capability - I still need to understand how it can coexist with other reporting tools

• A combination of rules, unified customer view, and a solid content hub can deliver can deliver personalization at scale (check the use case attached from Pandora)

• Some companies put Tealium at the center of their data ecosystem (check Twinset architecture attached) and with server-side tracking and omnichannel product recommendations were able to significantly increase Google Ads ROAS and conversions on Meta responsive audiences

• Some other companies have linked Tealium to Snowflake to prioritize calls to call centers, presented customized quotes to customers in their digital journeys and optimized the customer lifetime value in their paid marketing activities

• The combination of CAPI, zero- and first-party data, enriched with AI, can provide new and amazing outcomes. 

• Many companies talked about their Analytics Team Structure

  • Some companies adopt a central team (e.g. Nordea) while others a "hub and spoke" one (like Lloyds Bank). The first structure, which I think interesting but complex to implement, requires Business Owners to be included in the tracking implementation, and to be the final ones to approve production changes. The "hub and spoke" one, which in some ways we have in Sunrise, consists in having a central team focusing on the cross-business functionalities, while the spoke define division specific tagging.

  • In each structure, JavaScript programming skills are key for analytics implementation. Some structures put a deep focus on CDP (having dedicated CDP Lead, CDP Business Analysts and CDP Engineer positions).

  • The key of making a strong Analytics CoE is to assign to it ownership of attributes, events, audiences, processes, regulation, use-cases, budget and education.

• There was an interesting non-business discussion on AI and quantum computing by Prof. Hannah Fry, in which I particularly liked the dilemma of considering consciousness an emerging property of intelligence (we should then expect a conscious AI soon) or the consequence of million of years of evolution of life on Earth (in this case AI probably won't develop consciousness).

What is my take on all of this?  In Sunrise we're in a process of rebuilding our analytics team and restructuring our tools (Tealium in particular). We are still very early in this process, and we need to put a strong effort in data quality. But all of this provides valuable food for thought as we build our team and explore the scenarios we can unlock once we're up and running.