Customer data platforms (CDPs) are a hot topic in the digital marketing world, and for good reason. Customer data is gold, and if used correctly, provide countless ways for marketers to provide personalized customer experiences. Danny Abraha, Strategy Consultant at Lev, had a chance to talk to Martin Kihn — Senior Vice President of Strategy for Salesforce Marketing Cloud about his latest book, “Customer Data Platforms: Use People Data to Transform the Future of Marketing Engagement.”
Danny: DMPs or Data management platforms will become a dying product category and CDPs are described in your book as an evolution of the DMP, but the problem is in 2019 when senior level marketers and agencies were surveyed on what CDP they've used in the last 12 months they said Salesforce, Adobe, and Oracle. But at the time, none of those companies even had a customer data platform.
There's clearly some confusion around the topic, so for readers who may be uninformed, can you explain what a CDP is and how it differs from a data management platform?
Martin: There are technical differences and business differences — the main difference is that a DMP uses pseudonymous IDs and historically has relied heavily on the 3P cookie as the primary ID. Over time DMPs have evolved so they’re not all 3P cookie based, but they were designed for programmatic advertising. When I say pseudonymous, DMPs do not by design include any personally identifiable information — names, addresses, phone numbers, emails, even IP addresses for the most part. The second thing is on the business side — DMPs were designed for programmatic advertising which meant they had to support massive scale. What I mean by that is, a campaign could realistically have billions of impressions, DMPs were true big data, right at the beginning of big data. So the volume of information contained in a DMP dwarfed any typical CRM system.
CDPs on the other hand are designed first and foremost to contain information about customers and prospects — so it is usually personally identifiable information. It contains unencrypted names, emails, phone numbers to organize customer data. The second thing is that it is not primarily designed for advertising use cases. It is designed for marketing use cases — that would include things like email campaigns, web personalization, some social advertising and even just analytics about customers and customer segmentation.
Danny: So it sounds like they're different in the ways they're built and the business outcomes they produce. What type of organizations or industries would benefit the most from acquiring a customer data platform?
Martin: Well we’ve noticed that the RFPs come predominantly from, first of all, retail. As a category, retail is the number one consumer of CDPs. And now, consumer financial services and media companies, and it’s spreading. But those 3 categories are probably 75% of the users of CDPs. And when you think about the reasons 1) these are B2C companies and 2) they have a very large scale of customers — they aren't a niche. And they tend to be functioning in both the digital world and offline world, so a digital storefront and a real storefront. What that means is, they’ve accumulated a lot of different tech platforms, vendors, MarTech systems that have information about customers and prospects, so they are facing a situation where they have a lot of customer data that is disconnected and siloed. The business benefit of them joining it and unifying it is greater than in another industry where they have fewer channels or less customer information, or don’t rely so much on the competitive advantage of doing customer analytics. So the organizations that benefit the most are the ones that have the most confused customer data at the greatest scale.
Danny: That makes a lot of sense. So, what is the return on investment a company can expect, whether it’s retail or financial services , by unifying that disparate data useful for them? Besides getting a better understanding of the customer journey, where do they extract the most value from a CDP?
Martin: Think of it this way — a lot of companies have two different MarTech systems. Maybe one for email and social advertising or a loyalty program. If they’re disconnected, what they’re essentially doing is analysis on two different customers in two different situations and they’re coming up with insights based on the information that exists in that one system. If they can somehow put them together or join them, they have a profile that contains information from both. The insights they get are based on more information on the customer. They’ve essentially enriched a profile using data they already have, in the same way you would enrich a profile using 3rd party data.
So I have a customer list, I go to Experian and I get some information about their credit score, I now know more about those customers. Well, internally we can do that as well. f I have an email system and know people’s different reactions to campaigns, and I also have a loyalty system and I can know what tier they’re in and what kind of rewards they like to redeem. I can join those up together, I can enhance my marketing from insights from loyalty. It basically makes all the activities like segmentation, or next best action, or next best offer, or even just measurement — like what’s the ROI or lifetime value — more accurate. It can benefit companies incrementally, if they only join email and loyalty, they will get benefits right away. Then they can roll in e-commerce, roll in social insights, storefront data so it doesn’t have to all be at once.
Danny: In your book you write, "At its heart, the promise of the CDP is to go beyond solving for disparate data, and start putting customers at the center of every decision. Connected customer data would be able to predict demand for products, sense trends in the market, and adjust hiring based on data signals and adjust capital investment." I'm interested in some of these use cases. Could you give some examples on how this would be possible?
Martin: It’s called a customer data platform, and it’s customer centric and normally about building a profile about people, but it doesn’t only contain information around customers, it can also contain other sources of business information that is useful. It’s just storage, so you can have product information as an object that is relationally linked to these customer profiles, or you can add information around store locations. So there’s a lot of richness once you’ve got an understanding of your customer behavior and you’ve organized the customer data in a way that’s better than what you did before so you have a complete view of the customer, you can then start to learn things. For instance, you’re doing segmentation, and in the past you had some basic segmentation based on age. If you have a customer data platform you can start doing better machine learning segmentations and build clusters that aren’t obvious, but they’re groups of people, maybe even more of them. Then we can say, “Well, these type of people use this type of product.” I can do customer research now based on this new group I’ve discovered to determine what other products I can sell to this group, but I wouldn’t have known about the group at all if I hadn’t had a CDP. That’s one example.
You can also unlock trends about, for instance, location. Let’s say I have 20 different new segments and 10 of them only shop online, never enter the store, and 10 only shop in store and never online. This is an extreme example, so basically, that would have an implication on store design because the group that shops in store probably have different wants and needs, and different product preferences than the other group, but I would have never known that if I hadn’t identified these other segments. There’s a lot of complex or not so complex analytics that can be done just having more organized and better customer data. The real assumption about all of this is that you, as a marketer, as a company, are getting to know your buyers better and in some cases much better.
Danny: Where do strategy consultants fit into all of this? How can marketers make better use of segmentation?
Martin: That was interesting to me. We (co-author Chris O'Hara) did this book , and when Chris and I asked potential buyers and even users of CDP, “If you had a perfect single view of the customer, what would you do with it? What would you use it for?” And that’s not always obvious. They want it, but aren’t really sure what to do with it. The answer here is just being given access to the data to your team at some level and letting them come up with segmentation and figuring out how many people are in a potential group.
For example, I’m on the marketing team and I decide there’s a really good opportunity for high net worth individuals in the North East for whatever reason. But in the past, it’d be very difficult to know if entering that market was worthwhile or even worth the effort. But if you have a CDP, you can very easily drag and drop and say what’s the LTV and what’s the location. It seems trivial, but you’ve opened the door for a lot of ideas and hypotheses around analytics. I keep mentioning segmentation, but it’s really just counting — figuring out how big different groups are. And that’s a really great use case. If you want to get more fancy such as next best offer, next best action, etc., it will become more important as AI gets rolled out as companies get better at targeting people based on what we know about them. Predictive analytics based on the data in the CDP can then become very powerful.
To learn more about what Martin Kihn thinks about how marketers can effectively utilize data to power their marketing efforts, check out his session at Ultraviolet, "Data-Driven Marketing: A Conversation with Martin Kihn."