In this Research NXT interview, David Raab, the founder of the Customer Data Platform Institute, who is also a marketing technology expert, an independent consultant and author of many articles on marketing tech and analytics, talks about how AI and CDP together can unlock infinite possibilities for marketers in the present era. He emphasises how AI can empower marketers by aligning tech, data, and creativity to deliver personalised customer experience at scale. Additionally, at the end of our discussion, David shares his willingness to know how much AI marketers think they are using versus what they are using.
Key highlights
What this conversation covers
- The role of CDP in AI-enabled marketing.
- How AI is evolving to take up more specialised roles in marketing.
- The edge marketers get with AI-powered marketing automation.
We are seeing AI creep up the ladder of complexity and take over more complicated and coordinated tasks, much like human counterparts.
How have things transformed in the martech space so far, and how is AI taking over now? And what kind of changes can we expect soon?
AI has undoubtedly had a significant impact on marketing and technology; however, I do not think it has taken over. We see an evolution in AI from performing narrow tasks like predictive modelling to taking up more specialised applications that need to string together multiple tasks and take over more of a coordination and management role. AI in today's context has evolved towards managing the entire customer journey and not just one instance at a time. AI-enabled marketing systems will eventually decide and orchestrate the whole customer communications process independently and in an optimised manner at scale. So we are seeing AI creep up the ladder of complexity and take over more complicated and coordinated tasks, much like human counterparts.
What is a real CDP, and how does it complement an AI-enabled environment?
The point of the RealCDP programme is to establish a universal standard for all systems that call themselves CDPs, so that customers across the globe could have a defined set of expectations of them. Everyone has an intuitive understanding of what a CDP is; however, it is not well articulated, so all systems that call themselves CDPs do not meet those expectations. The fundamental purpose of the RealCDP programme is to help people make the CDP buying decision more confidently.
The actual requirement of a true CDP is that it should be able to intake data from all sources, store that data indefinitely, present the data in unified customer profiles, and it should also be able to share the data with all the systems. This is a basic set of requirements for a CDP, which systems like a DMP or CRM do not provide in one way or another. A true CDP should have all these capabilities, even if it is through an API or if it requires manual entry by the vendor or the user to expose the data to other systems. The reason these requirements should be in place is that people should get the CDP to do what they implemented the CDP to do.
What are the top AI in marketing use cases wherein CDP would contribute?
The most important thing with AI is the data it is trained on, and the core role of a CDP is to assemble all that data and make it available in a format that is suitable for the application, so that AI can work on it. Hence, any AI in marketing use case would benefit from a CDP. In terms of specific use cases, we have seen predictive modelling and campaign designing most commonly. In terms of chatbots, CDP is useful where customer data is used to benefit the call centre. Agents take the next best option recommendations powered by AI.
What should be the trigger for an organisation starting to consider a CDP for themselves, and how should they go about evaluating one?
CDP is ideally for an organisation of considerable size. They are usually for mid-tier or enterprise companies where there is enough data and complexity and enough resources to take advantage of the CDP. In terms of industries, initially it was primarily in retail and in online media where you have a lot of frequent small transactions per customer, where there are many recommendations to be made. In scenarios like this, it is easy to measure the value added by a CDP. More recently, we are seeing CDPs being used pretty much everywhere, like services, transportation and hospitality, and beginning to see it in education and telecom too. Another benefit of CDP, apart from the financial benefit, is enhanced CX through personalisation at scale. And since every customer today has high expectations, CDP is useful across industries, across company size, and business models.
At what phase of an organisation should they be starting to implement AI for marketing?
In my opinion, AI is productised, and simplified AI is already in use across organisations. I do not think AI needs an organisation to reach a particular stage of maturity to be implemented. It does not mean organisations need to create AI from scratch, as more of the strategic systems like CRM and marketing automation already have some AI embedded in them. I think everybody has access to AI, and they can take advantage of it without being very technically adept. It does not mean that marketers should blindly trust AI. They should check out what the AI is good at and how it should be applied appropriately, so that the data they need is there and can be reasonably used.
How was your recent experience in India while you collaborated with Netcore to host the CDP workshop? What did you sense about the current progress of martech in India while you interacted with marketing leaders here?
I like India. India is a complicated space with a lot of smart and educated people. The challenge of doing business in India is very different from other places. A vast section of consumers in India has access to mobile connectivity irrespective of their economic background, and it is fascinating. We do see great technologies in India; we see companies having a lot of different use cases. For example, one use case was that of a bank that had a hard time figuring out the financial status of its customers as the data was not readily available, unlike in the USA where it is pretty easy to access. So before anything else, customer data needs to be in place for technology to act upon it.
As a marketer, what should be on a checklist while evaluating or upgrading the systems to a CDP?
I think marketers need to understand their situation and create their checklist based on what use cases they care about. They will also need to identify what specific requirements they have from a CDP. Is it just the customer data, or do they also wish the CDP to do the analytics and also run the marketing campaigns? Some CDPs do that and some do not. They will also need to choose which channels for outbound the CDP should support. The choice depends at a high level on what scope of CDP functions the buyer wants in their particular situation. Real-time recommendation is another function that should be considered. If it is needed, buyers need to understand how the CDP makes these recommendations.
And finally, what is the one question that you would like to ask marketers across the world about AI?
There are many, but what comes to my mind currently is: how many of the marketing systems that they use have AI capability built in? I am curious to see the answers, and it would be great to compare them to reality. It would be interesting to see how much AI marketers think they are using versus what they are using.
