Getting started with a Customer Data Platform (CDP) can be confusing since there are so many things to consider and competing claims out there. This is something we think about all the time: how to help companies find their way to value through a complex and often confusing landscape.
BCG recently wrote a useful overview of how CDPs can help achieve better personalization and what companies should consider when thinking about a CDP initiative. We had the chance to ask a few questions to Stefano Fanfarillo, Partner and Director of Personalization and Digital Marketing at BCG and one of the authors of the guide.
(Amperity) What’s your simplest definition of a Customer Data Platform?
(Stefano Fanfarillo) I think of a CDP as the technologies and processes to support a marketing operation in four key areas:
Collection and integration of customer data across multiple sources
Creation and maintenance of the 360-degree view of the customer
Enabling the development of relevant customer insight
Activation and orchestration of audiences across multiple channels and media
I include all four of these points because I believe it’s important to emphasize the breadth of capabilities as well as the necessity of the whole process.
In your paper you identify three obstacles to 1-to-1 personalization that CDPs can help address: disconnected data, lack of advanced analytics expertise, and lack of marketing technology integration. Would you say there is one that stands out as the most persistently problematic?
We generally find varying degrees of all three challenges across all the organizations we work with. I’d say the most consistent one is the disconnected data. It is an absolute challenge for any organization, or technology vendor, to be able to keep up with the explosion of volume and variety of data that exists in the ecosystem as well as keeping up with the pace of change and new types and sources of data regularly appearing – e.g. new products, interaction points, social platforms, and so on.
You wrote that building a CDP in-house takes a lot of time and resources. Why might some companies choose to go that route?
There are several reasons why an organization may think about building their own CDP as opposed to using an outside vendor. Some of the main reasons I have come across are:
The organization has unique solution requirements and doesn't believe that off-the-shelf packages can deliver everything they need.
Many organizations now see data as a strategic asset and may feel uncomfortable with handing their data to an outside vendor.
Some organizations believe they can do it more cost effectively themselves.
Finally, a very common reason is when an organization already has something that can deliver most of the capabilities (e.g. a data lake) and just needs to add a little to it.
Are time and cost the main considerations you see for not building a CDP in-house, or is efficacy an issue too?
A key recommendation for any organization looking at a buy vs. build option is to consider the end-to-end spectrum of capabilities and what the end goal may look like. Looking at a CDP only from a data & analytics perspective often leads an organization to fall short of what they really need. There are several aspects an organization should consider in making this decision:
Required capabilities: do you have the skills and resources to build a CDP? Do you have the resources to maintain it? Sometimes it can look doable to build but it is then hard to maintain.
Speed to value: what is the quickest option to get to an outcome and what is the opportunity cost? Perhaps it may be cheaper overall to build, but what if the buy option takes you to a position to deliver value quicker?
Ability to evolve the solution: will you have the required skills and resources to continue to evolve the solution as new data sources appear, new interaction channels appear, new use cases come up, and new marketing tools appear in the stack and need to be integrated?
User experience: you need to have a solution which can be operated by a marketing user and not an IT persona or a data scientist. Off-the-shelf CDPs have been developed with the marketing user in mind – one of the key use cases is to be able to expose complex data in a simple and usable way.
Implementation approach: a CDP must be adopted by the marketing user and be embedded into the marketing operation process. "Build" sometimes becomes an 'IT engineering' challenge with a lot of time and effort spent on solving for that technology layer, as opposed to a marketing operation improvement challenge, which is the main goal of deploying a CDP to start with. A “Buy” option has core functions already built which removes some of the engineering effort and likely makes it easier to adopt an agile and incremental delivery model.
If an organization doesn’t take all these considerations into account from the start, it risks creating a solution which may not be able to scale or deliver all the capabilities required. Most importantly it may end up with a solution which is hard to maintain and requires highly specialized resources. It may go from being a ‘core enabler’ to a ‘key obstacle’ of marketing operation agility.
You advise companies to kick off their CDP initiative with a specific set of use cases in mind (and we agree wholeheartedly). Have you found advantages to prioritizing one type of use case over another?
When we work with our clients, we want to start with a discrete set of use cases with the goal of delivering value as quickly as possible. The use cases will typically involve end-to-end execution, from data to analytics to activation to measurement, and they are focused on a minimum viable product that can be developed quickly.
For example, one apparel company we work with wanted to deliver highly personalized offers across its owned channels, including email and websites. A relatively simple use case like this, if it meets the expected goals, can be deployed fairly quickly and build momentum across the enterprise.
In order to deploy an end-to-end use case, like the example listed above, all of a CDP’s capabilities are required: ingestion & unification, intelligence & decision making, and activation. However, as use-case sophistication is incrementally built, along with the incremental value delivery that comes with it, the level of sophistication and maturity across the CDP’s different capabilities is also incrementally developed. In this more incremental method, companies unlock value as they mature and scale their personalization efforts—and can fund a long-term transformation out of the returns from the use cases.
As companies build up analytics expertise through using a CDP, how important is it for them to work with outside guidance like partners at a CDP vendor or a consultancy?
It is very important to talk to the experts – any of us would want to do that for any job and it is even more important when doing something as complex as a CDP. The ability for any organization to organize, analyze and leverage its own first party data has become a core competency and, in many cases, a key differentiator in the digital economy. Organizations should look at outside experts to help understand what the long-term capabilities should look like, what may make sense to build or buy and to get a deeper and independent understanding of the vendor landscape and which vendor would be best suited to support specific use cases and support the long-term ambition.
Can you share a case or two where you saw a CDP really help a company succeed?
I have worked with many clients over the years helping them implement, in some shape or form, CDP capabilities – whether custom built, deploying a vendor solution or a mix of the two. CDP capabilities, in our broad definition, are a foundational element of any data driven marketing initiative and therefore a must-do for any organization wanting to drive a strategic personalization agenda. Following our use-case-led, outcome-driven, incremental delivery approach, they have all been successful.
I’d say perhaps it is more about the different degrees of CDP capability sophistication and maturity we have been able to achieve across multiple client engagements. However, this is fine too because not every company needs to get to a ‘best in class’ level of maturity in order to deliver the desired value. In fact, the goal of our stepped approach is to avoid the redundant features, extra integration work, customization costs, training, and overall uncertainty that plague IT initiatives in general and MarTech implementations in particular.
Download our 1-page quick guide to getting started with a CDP for a handy reference of key things to keep in mind.