The Martechno Beat: David Raab on the Role of CDPs in Delivering Personalization at Scale



The ongoing COVID-19 global pandemic has confined most of us to our homes through government-sanctioned lockdowns.

In such a scenario; customer behaviour, consumption patterns, and spending habits are witnessing a shift, across industries and geographies.

A central component for growth even during these unprecedented times is customer data unification, cross-functional sharing, analytics, and actionable insight generation.

This lays the foundation for the rising evaluation and adoption of Customer Data Platforms (CDP).

In fact, in 2018, industry revenue reached $740 million, up more than 50% from the previous year. CDP also featured prominently on Gartner’s Hype Cycle for Digital Marketing & Advertising, 2019.

All this points to a reality that the concept of CDP has gone from being merely a buzzword to a need of the hour for businesses across industries.

So, I figured I’d catch up with none other than David Raab – the Godfather of CDP and the Founder of the CDP Institute – himself, as part of the current season of India’s first-ever martech podcast!

Read on to hear what David had to say on all things CDP and how it helps B2C brands fuel personalization at scale…

PH. In a lot of my professional conversations, I’ve realized that the definition of a CDP often varies from marketer to marketer and across industries. How would you describe it to a 5-year-old and then to a grown up?!

DR. To a 5-year-old, we would describe a CDP in terms of different kinds of vehicles.

A CDP is like an ice-cream truck as it has a lots of different components with different people within a company chasing after it.

They are delighted to see it and want to pull different kinds of data (read purchase ice-cream flavours!) from it.

This is opposed to a data warehouse which is more like a schoolbus and a data lake is more like a dump truck!

In more grown up terms: A CDP pulls customer data from different sources where it sits or gathers and puts it in one place. This is the kind of data that doesn’t usually talk to each other. Businesses can use it to get a complete picture of individual customers.

A CDP then shares that information with any other system within an organization that may then want to use it. Consider CDP as packaged software that you can typically buy and deploy more quickly; unlike data warehouses that need to be built from scratch. It is flexible in terms of ingesting data from any source and pushing out data to any source.

PH. How much time does it take to implement a CDP from scratch?

DR. Actual installation takes a matter of days. The implementation takes longer as you have to decide what sources of data you want to feed into the CDP and assess the quality of this data as well.

You may require a certain amount of data modelling, but that depends on the CDP. So, realistically, a business should get a CDP up and running anywhere between 6 months to 1 year. That’s far less time when compared to building out a data warehouse that can take 2-3 years.

PH: You’re also accounting for the degree and level of integrations required for the CDP to be ingesting that sort of relevant data?

DR: Right, again. Most of the CDPs have pre-built connectors for the most common data sources. So, they’ll connect to your e-commerce platforms, sales automation systems, etc. That’s just a question of pushing a button. But, if you have to build a custom connector, that will take days or weeks, but not months.

PH. Who would be the right personas or key stakeholders to be driving the implementation of a CDP at a business?

DR. We see a few different personas involved. The most common buyers are the Director-level marketing people such as the Head of Digital/Email/Advertising.

Sometimes there’s a marketing technology staff that reports directly to the CMO that might drive this project. We’re seeing more CXOs getting involved such as the Chief Data/Analytics/Technology/Information Officer who wants to drive a corporate-level project, and not a marketing-level project.

We see continued and expanded interest in use of CDPs throughout the organization. And, with new data regulations in place, we also see the compliance teams get involved to ensure that customer data is handled correctly.

PH. We’ve now entered 2020. CDP has emerged as a key trend over the last couple of years. What are some of the industries in which you see rising adoption?

DR. Adoption is actually rising across all industries. Initial adoption was clustered within the retail, media, and publishing industries. This was because these are industries that see a large number of frequent low-volume transactions per customer, so there’s a lot of data that can be put to use for segmentation and personalization.

Over the last couple of years, adoption is rising across the financial services, insurance, healthcare, travel, and telecom industries. While these industries witness fewer large transactions but increasing volumes of data being generated – this customer data can be used for personalization, even if it doesn’t result in an immediate purchase.

PH. It’s interesting that you mentioned the rising adoption in the financial services industry. We’ve got a lot of banking customers who are still using data lakes. In a situation like that, how helpful would a CDP be?

DR. A CDP will often use a data lake as an input source. The idea of pulling this data together to create a unified view of each customer requires merging identities across different systems. That’s not something a data lake can do. It simply provides a place for the data scientists to pull data in its raw form for the purpose of analysis.

A CDP basically normalises this process and does away with such manual intervention, executed in the background so other systems can use that already prepared data.

Data lakes and CDPs are complementary to each other, in most cases.

PH. We spoke about the benefits that a CDP bestows. What would be some examples of companies that you believe are leveraging CDPs optimally? Particularly when it comes to powering personalization at scale.

DR. We at the CDPI look at what different vendors are doing with different customers. Barkbox, an online food subscription, is using a CDP to assemble a unified customer view and segment subscribers who would cancel subscriptions. They then used these insights to run experiments – in terms of messaging, promotions, and campaign delivery times. This approach helped them 2X the revenue from that segment.

HelloFresh used a CDP to optimize their direct mail and email campaigns and ended up increasing conversions by 2X. Advantage Rent a Car leveraged a CDP to drive dynamic personalization of their home page based on who the visitor was, where they came from, and what they looked at before. This increased bookings by 25%.

Good companies use CDPs for many use cases as it is a foundational technology.

Once you have those unified customer views in place, the number of applications can be infinite.

PH. You mentioned how Advantage Rent a Car was dynamically personalizing their website. The likes of Amazon and Netflix rely heavily on their recommendation engines. Would that also be a viable use case?

DR. Yes, absolutely. One of the primary use cases is to support analytics and predictive modelling. Anyone who spends time in data science, spends most of their time in data preparation and that’s what the CDP productizes. So, once that data preparation is done, it feeds into any kind of recommendation engine or predictive algorithm.  

PH. Everyone’s talking about CDPs, but a lot of marketers think that it is not for them or it is for their neighbours! What advice would you give to such marketers?

PH. I’d first like to ask them why they would think such a thing! We think a CDP is useful for most businesses. There are some practical constraints pertaining to small businesses. The core value is that it brings data together from different systems, so if you only have one system that has all your data, then you may not need a CDP.

You also need a certain level of sophistication – in terms of marketing intelligence, infrastructure, martech stack, etc. – once you get past the basic use cases to tackle more complicated use cases like personalization. Not all companies have those resources.

But, if you think a CDP is for your neighbour, then you should go work for your neighbour as they are likely to thrive and your company is more liable suffer!

PH. We see a lot of online startups that are happy to adopt a multi-channel marketing hub or mobile marketing platform. So, they don’t have a standalone CDP involved in the entire equation. How do you look at that trend?

DR. If they’re only online, they’re present on one channel. So, having a standalone CDP may not be the best idea. But, if you’re using a multi-channel marketing hub – it usually has some form of a CDP within itself along with the delivery systems. So, if you have just the one platform (such as a multi-channel marketing hub) and that’s working well for you, then you don’t need a CDP.

But, if you’re using separate email marketing, web personalization tools, etc. then these multi-channel marketing hubs may not have the setup to stitch these data sources together and enable cross-company sharing. That’s when a CDP would add value. So, we expect to see more adoption as the reality is that very few companies can find a single vendor that can do everything that they need as well as they need it to.

So, we expect to see more adoption as the reality is that very few companies can find a single vendor that can do everything that they need as well as they need it to.

PH. That brings me to my favourite question! Where does the future lie for CDPs? How do you see this space evolving?

DR. This is a very dynamic space.

CDP is expanding beyond its definition; i.e. A class of packaged software that builds unified, persistent customer database sharable with other systems.

What we are seeing is that there are some true CDP systems that merely build the customer database. And, then there are other systems that offer CDP, analytical capabilities, journey orchestration, and/or delivery mechanisms.

We are also seeing more vertical specializations. For instance, there are CDPs that cater to the travel, healthcare, or retail industries. They have connectors into the operational systems – ticketing systems for travel, medical records for healthcare, etc.

All this confuses buyers. So, the reality is that once you know your industry, needs, scale, budget, and scope of functionality; it narrows down the choices. But, we’ve gone from businesses not recognising the need for CDPs 5 years ago to now having hundreds of options to choose from. That’s a second order problem, and a far better one to have!

Check out the entire episode (and gain access to some martech banter!) right here and stay tuned for more such insightful content!

Join 20,000 Marketers who get our Blog Emails Every Week!

Kickstart YOUR growth journey with our expert tips on marketing automation, personalization, email, mobile marketing and more. Only 2 emails per week.

Pradyut Hande

Pradyut Hande

A data-driven marketing strategist with a hard-nosed penchant for frugal innovation, Pradyut is the Senior Growth Marketer & Product Evangelist at Netcore's Smartech.

Join 20,000 Marketers who get our Blog Emails Every Week!

Most Popular Blogs