Santosh (Host): Hello, viewers, listeners. My name is Santosh Valecha, and I invite you to get another episode of a fantastic podcast from Netcore’s table. Today’s session is special and we’re going to be talking about a very hot topic, the buzzword, the keyword, something that everyone’s looking for is CDP, also known as a customer data platform. And to talk about the subject today, we have with us someone, a very dear friend of mine who I’ve known for the last decade or so. He comes with over 12 years of experience in the field of analytics. He’s done it all; descriptive, descriptive, predictive. He’s been there, done that. He’s helped me create a lot of models in the past. And I’m excited to bring to you none other than Kunal Sakaria and Kunal, welcome to the podcast. I’m really happy to be connecting with you virtually, but it’s a pleasure to have you with us talking about this special subject. I know you specialize in this and viewers will be excited to know that Kunal also drives the CDP product experience overall at netcore. And I’ve invited him here today to give us some background to what you do for us. But I’m happy to have you.
Kunal (Guest): Hi Santosh, it’s a pleasure to be part of this interesting topic. Discussion on CDP, I think.
Santosh: So, many times I’ve been talking to a few friends of mine in the industry and, you know, and there is always this question, what is CDP, a buzzword? A lot of people are talking about what CDP is. And, you know, it’s an industry trend, if you want to simplify it for our listeners. In your words, what is CDP for you?
Kunal: Yeah, I mean, so I suppose to expand on the abbreviation, so it stands for Customer Data Platform and a lot of marketers have been hearing about this storm recently. What do you think? In fact, for the last five years, but I think recently it has been catching up a lot because a lot of these have now started off putting CDP platforms. So, I think they will probably be go-to players. But now, as we stand today, there are over a hundred plus CDP players in the market. So that’s the kind of interest that it has gathered in the coming years. And so basically, CDP is a marketing platform that collects data from various customer-facing touchpoints or channels. And then it unifies all those data points to create a single, accessible, and Real-Time view of the customer.
Santosh: OK, so we think that we’re going to get one view of the customer or a unified view of the customer.
Kunal: Right. How is it different? Because the unified view is not new, brands have been trying to do this since, you know, a decade or so since we had the emergence of a lot of marketing channels. Earlier, there were only SMS & emails. But now, as the channels have emerged, marketers have strongly felt that need for, you know, having a unified view of all their customers since all this by law. But what has stopped them from this is all these data points are lying in silos everywhere. And a lot of marketers have been struggling to do that, and hence CDP has a product platform and has seen a lot of traction in this in recent years.
Santosh: I actually agree completely agree. I mean, I’ve been a marketer myself. I know that we’ve always had data stashed in different places. And bringing them together is such a difficult task for the different platforms. And there’s center data sitting there. There’s digital data somewhere. There’s unstructured data, unstructured data, transaction data that the customer profiles to be somewhere else. I mean, there are so many different ways in which you can bring data together. There’s a really interesting thing that people talk about and the keyword that is always being is CRM. And then there’s another three-digit word, which is a DMP. How do you differentiate CDP from these?
Kunal: Yeah, so before I go there, it’s like I was saying, the unified view is not a new term you know, it’s always been there. But what differentiates CDP from a unified, stand-alone, unified solution is that CDP can also help you act on these unified insights. That’s one key difference that a platform like CDP brings to the table, which is acting as your brain or your data, getting all those insights and also seamlessly helping you to action on those insights because in a CDP, marketing automation platforms, or it could be email delivery engine or it would be like a cookie retargeting platform. This may make all these New Age things or you can be an automation platform. They’re all bundled inside this package. So hence it is an extension of the unified view concept. So that is one key difference the CDP has got. And I like to ask you, what is the difference between CDP vis a vis a platform like CRM or DMP? So, the key difference here is that CDP works on first-party data and third-party data. So, first-party meaning brands, all data which collects about the customers to their own properties, which could be a website app. Or let’s say some e-mail communications to their customers getting back that response data, or it could be causing the communications. So, all that data belongs to that brand and hence that becomes the first-party data. And the second is third-party data, which is, let’s say, when you would tie up as a brand with an external platform for, you know, running cookie-based campaigns to target you, target your customers on other platforms. Or let’s say you don’t get their profile’s behavior in terms of what kind of websites they like to visit? So, all that becomes third-party data, so that can also be ingested by CDP. Whereas if you look at DMP, it’s primarily focused only on the third-party data and for CDP, it is just another source, but then CDP extends beyond that. It is not just limited to third-party data. It’s in data from as many channels also as you can think of, as long as it is in the realm of structured and, say, structured data, unstructured data kept aside because a lot of CDPs are still building those capabilities to ingest audio images, audio data or images, data that is still far off. But leaving that aside, any other data can be directly injected into CDP through connections. So that is one. And if you were to compare it with the CRM platform, the key difference there is that in CRM it again focuses mainly on first party data. You know, that your time action logs or those conversations that happen. So, it has a perfect track of all those events. But what it doesn’t have is multi-capability. So, if you see, I cannot tell you what your customer is likely to do next, whether he is going to repurchase the next month or what is the next brand or the product that is likely to buy some e-commerce retail website or do that kind of a business for CRM that lacks those capabilities. Also, let’s see, I cannot give you insights on the Web trail of the customer. So, once he’s on your website, you buy sections of this website. How much time is he spending there and what kind of links that you click on? All that is not drag by CRM, whereas CDP can crunch all those insights and value that your customer was actually browsing on your website and app and this is what he looked at and this is that understanding, you can also go to recommend the next best product, or it can tell you when the customer is likely to re-purchase from you. So, that’s how we differentiate from CRM. So, it is definitely enabling a lot of use cases, which traditionally CRM could not do. And at the same time, you cannot get from only a unified resolution, which I choose to have or without the action capabilities, so hence it becomes a complete packaged solution to take care of the entire customer’s cycle from his, say, anonymous voting activity. To the point where he becomes an identified customer for you and then also tracking what he is doing on your platform even when he is not transacting.
Santosh: So, it’s not just data, it is intelligent data. It is data that has been culminated by bringing multiple data sources together in your experience, building the entire suite of offices. What are some of the interesting use cases you think our marketers can leverage using a CDP?
Kunal: Yeah, so I think in my experience, of course, has been a new development, but we have deployed this for quite a number of clients now and we’ve also been speaking to a lot of clients in and outside of Indian markets as well, including Southeast Asia, the Gulf region. And so, this is what we’ve seen is marketers typically first want to create a unified profile of their customer by switching the identifiers across multiple devices, multiple channels, so that is one key use case because otherwise what happens is the same customer is duplicating. So, if a customer profile is defined in terms of a device or it could be a mobile number. Now you might have multiple variations of that. Great customers can use multiple devices. You might have multiple mobile numbers, multiple similarities, and so on. So, there is this abundance of identifiers. But then for brands, they don’t want to communicate each and every time. One that identifies. You know, they don’t like the fact that it’s the same customer and they want to do that, so hence stitching the profile of the customer to create a unique profile. I like how that phone call works. It’s a unique profile that is created as an output. So that is one use case that enables a lot of relevant targeting that the brand can do to their customers. And it is not just, you know, communication. That is, you lost it without doing that. The same customer is receiving it multiple times. That is one use case. Basically, it brings you to the entire communication plan of the engagement plan, which even I would have. And then when you do that, the customer realizes that, OK, this brand is not just bothering me with emails or unless a message or notification without doing that, I’m the same customer. And when you do that, the customer sees value in your communication. So that is one use case for which marketers are using this platform. Second is, it provides you access to all of these data points, which traditionally marketers didn’t have access to. So, if you think of data warehouses or data models that existed, they were never made accessible to marketing teams. As the marketing team would have a lot of use cases around it in terms of, let’s say, they would want to create a segment of customers. Who have they bought a product in the past, which is a lookalike of the product, that they’re going to launch the future, and when they wanted to let you do this and will use case where they wanted to simply target customers who are lookalikes of these customers who have bought this product in the past, then they would have to go to that to get this data. And IT would then again take a turnaround of a couple of days to crunch that list with the required segmentation rules and give it back to the marketing team for running campaigns. And then, of course, there would be a campaign execution which would pass on this data to a campaign automation platform or to any delivery partner, you know, depending on the channel on which you want to reach out. So no, CDP does not require all of that. That’s completely eliminated that entire long-time cycle of exchanging data in various needs. So, it has democratized what happens if no marketer themselves can have access to all of this data and they can create their segmentation rules on the fly. So, you just mean the audience that you’re looking at for reaching out to a particular candidate, which would be on a product launch or it could be a personalized offer that you want to say all those use cases can be enabled on the flight. You don’t need to go to an IT team. You can do the segments and get them.
Santosh: Kunal, what a big change that we’ve seen over the last couple of years is that there was a time when consumers had limited buying options or avenues to purchase. Today, even if I have to purchase anything, I have the option to go offline. I have the option to go online, have the option to go to the web, the option to go to the app. There are so many different ecosystems that have been built right now. What’s your take on this? How does it bring the centerpiece system together? Because my behavior offline is very different, maybe even online two very different things I buy when I go to the store could be very different. Online has made it so much easier to buy cheap items at regular intervals of behavior so different from a customer point of view that takes into consideration the singularities.
Kunal: Yes, yes, in fact, that has actually led to the CDP platforms coming in the first place. So if you see earlier, we didn’t have so many marketing channels, but now as the channels are going up, that customer experience has become fragmented. So, the example which you gave, right, you also up from the website and then you might also sometimes go into a store when you walk into a store, does that representative know that Santosh is the same customer who has also led to what a tremendous amount of has done a tremendous number of transactions on the Web site, and this is something that he likes to buy. And then this is that, let’s say, in the store he’s able to recommend new product options based on your brand product category preferences. It never happens, but that is what our objective is, to be releasing data from multiple channels. And no matter where you are, you know, whether your own website is at or in-store, the idea is to enable all those insights and. Give those insights good customer touchpoints, which could be a store representative or could be a website, which is a website notification or an app notification so that the entire experience changes for you. So, which is what I think CDP really meant to do, which is take all this segmented data, combine it into a unified insight and better the customer experience.
Santosh: I mean as a marketer, how do I know if I’m ready to bring a CDP of policy or subscribe to an offering that can enable my CDP? How do I know as a marketer?
Kunal: Yes, that’s a very good question. In fact, you know, just because the boss told me it doesn’t mean that everybody needs to take a plunge. So, I have a checklist here which I usually talk about lines that can help and help a brand evaluate whether they really need to embark on this journey or not. So, it starts with these five questions that I have. OK, so the first one is whether the brand is having multiple products or brand lines that are marketed to their customers. And that is the first thing. The second is, does the brand have multiple customer ordering channels, which will be your website or your offline stores. They become the ordering channels and see whether they have multiple channels or, let’s say customers or be in touchpoints where they get to capture the behavior of our customers’ response to the marketing campaigns. So that’s the topic. And then the fourth is whether the brand is looking for the right insights and analysis from a single location and the fifth is to what extent does the brand need to democratize data then, you know, how many teams are able to access these unified data points?
Santosh: OK, that was going to be my follow up question to you Kunal, so, ok, so I have gone through my checklist, have ensured and now who’s going to be the user of my physical platform with the key stakeholders who should be involved in the process of building or consumption.
Kunal: Right, so, of course, we need a champion from the organization level as the marketing team first senses the need for CDP. So, they usually, you know, get into the phase of evaluating CDP partners in the first place. And then after that, you know, once the CDP is able to solve those marketing use cases, that is when the other team starts jumping in, which is your science team, or it could be the IT team and even product team in fact, so it’s not just marketing team, which is a misconception that only they use when they are the first ones to actually bring this up in an organization, let’s say, on the department. But then CDP is also used by, like I said, product teams and IT teams to enable their own use cases. Also, if you ask me why each of these other things has come up, if you look at it recently, I had challenges tracking the entire trail of incoming and outgoing data. They’re always the first ones to see that data privacy is important and security is important. So, they want total control over the entire incoming and outgoing data, so a CDP has mechanisms to help you create that trail of data coming in and where it is going so that then the audit process becomes easier. So that is the one to use case by which I have also started hoping for. Then the team is, like I said, product marketing teams, alternatives, basically, that they do a lot of these tests and experiments on your website or an app to do those experiments earlier. It would take a developer test and validate those experiments. So, you change something on the website and now you want to see which homepage will be better compared to a couple of options and all those kinds of things. But with all that becomes very simple because you get those insights quickly on the fly. Whenever you do any change and all those recommendations can be enabled on the fly. Hence, product teams have also started adopting this for their use cases because it minimizes the entire cycle of doing those experiments. And the team, like I said, is the data science team. Now, these guys have always been wishing for a lot of data points. Well, you see, they’re always hungry for data, though, because I see myself being an analyst, we always have that complaint saying that the organization is not collecting enough data for us to win one of the best motors on. And hence and also, they are not the people who will spend time collecting data from various sources. You don’t want to get to do that. You want your data scientist to build your email models. So with their expertise or statistics and machine learning, which is what this does for sleepy people, you know, by virtue of being able to unify all the data in one place, it removes that headache of getting data together. And the job for the data scientist team then becomes easy. They just need to get into the model deployment phase. They can pull all that data into their analytics data warehouse or in some cases where there is advanced CDP which gives you these email tools or, you know, statistical tools to be run out of the box on the CDP platform. So hence the cycle audience or all of these teams. So, yeah, these are the themes in a nutshell.
Santosh: The two-interesting things about analytics I think are the two key areas that excited me the most is the fact that I’m able to build a repertoire of data from multiple sources into one place, which then also allows me to define the potential lifetime value of the customer. This is another area that the analytics really comes through because they’re so juiced up with the information sitting in the repository that there are so many things that they could do in a way that is interesting, literally. The second question that it brings me to is how the free market is in deciding how they’ve made the desired investment to build the CDP. They’ve got the right stakeholders because they’ve been chugging together. How do you define the ROI coming from CDP? How do I know that I’ve made the right investment at the right place?
Kunal: You know, that’s an interesting question, I think, where the buck stops there. So, I think if you look at it the way the brand can go about justifying its ROI through a platform like CDP it starts with IT cost. Now, let’s say if a brand were to get these unified insights in house with their own IT team capability, one is a lot of time because they are not experts in that area. Right. They are good with, you know, use cases or take around their daily activities in business. So, they’re not so well versed with CDP, the platform. So, when you ask them to build something like this internally, it is because they are not justified in that. In that case, there is a long-time cycle that you’re looking at, probably the range of things we do.
Santosh: Well, I’m sure there are some specialists out there who can really do this. So if you’re out there, then this is not true. One of the ways that many of us can subscribe to building a facility outside.
Kunal: If you see it that way, I mean, imagine the cost a brand would incur for the next 12 to 18 months if they were to do that. And so, they need to look at that cost and see that if we go to like, say, outsource this to an external partner and there are plenty of them in the market, like I said, 100+ players, each with a unique set of features and capabilities. So, brand can you evaluate if you are outsourcing to an external partner, what is the time to go? And what are the benefits that will accrue when they are able to go life in the next three months or a couple of months, if they go on with an extended part of their benefits, that taking the cost there would not be so high as compared to when you to build inhouse, ok, so that’s how brand can try to evaluate and see the cost. These are really benefits that can accrue to them in the immediate near-term. You know, they don’t need to wait for 18 months to embark on this journey. So that’s all they can look at this and say, again, if you see a lot of these CDP are also packaged with algorithms which help you sort of use cases and cross-sell upset. And if there is a brand, you can actually find a lot of these case studies, you know, by attending some of these CDP workshops or, you know, by visiting some of these websites, we’ll have these case studies where they can get a sense of what is the kind of uplift we can get to a CDP, if they were to run those whose cases on the platform, what is the uplift that they can expect and then even that can be quantified into a number which can be added to that list of benefits. And then we can then look at the savings that it does in terms of disposing of it, the data management cost and against that ROI benefits. In terms of incremental campaign uplift on across upscale retention and then basis that you can take a call that, you know, whether it really justifies to go for CDP externally or whether to go for it or not in the first place globally.
Santosh: And we can’t miss the fact that the experience, multi use experience of the partner system to bring to the table, which we may not have internally as a marketer to build, may have been given within some Minato. So, I think it’s very interesting that we do recommend that it’s better to be outsourced than to be built internally.
Kunal: Very valid point that you brought up the cross-industry experience which an external partner brings to the table. I think that goes without saying. I mean, it’s not just about what is right for your business. These partners externally bring new insights from across various other verticals where they have worked with so that you can get some heartbreaking insights on your ML model, which is great for a particular vertical, which might also be suitable for all those industry insights. The repository is a key asset for these companies and which is where the entire expertise lies. And that makes a good point for interviewing for an outsourced kind of a model of trying to develop. So, there are a lot of benefits, actually, if you look at it from a longer-term perspective.
Santosh: So, I can build CDP in any business vertical, I could be a travel portal, I could be a movie portal and it could be an aggregator to cut across all segments the ability to pull the CDP.
Kunal: Yes, definitely. And in fact, each of these verticals would have they’re place in CDP versions. So, if you’re an e-comm or a retail, the kind of CDP templates in terms of out-of-the-box segments that you want, I would differ from and legacy of a brand brown-nosing to insurance and the frequency of purchases, not that high, you normally just review your insurance policy once a year with e-commerce who might transact every month so, that’s a key difference. So, CDP has all these templates which are anybody for each vertical and those are tailor-made for each vertical, which helps you to achieve use cases specific to your vertical. OK, so let’s say if you are insurance, the CDP has a male model that will predict the propensity to renew whether the customer will renew its policy or not. Right. And if you are in e-commerce, it will have an ML model that will create predictions whether the customer will repurchase or not in the next one. Right. So that’s all of these templates. And they are tailor-made for every industry so that they use cases that are not. But then again, it differs from CDP to CDP, so not all may be so vertical centric, so maybe generic CDP versions as well. So that is what a brand can actually question, when they are getting on board and getting on board CDP, these are some of the questions to ask.
Santosh: Is this one last question that most listeners and viewers would want to know is how do I go about choosing the right CDP platform if what would be your tips to market itself out of the group or choosing one?
Kunal: Yes, yes, in fact, on this, like I said before, the use cases that your business wants to, its which is not just marketing all sit in a room, big marketing stakeholders, your product stakeholders and also your data scientists can have the list of use cases that they want to enable. This unification process of unification is not the objective of CDP. It’s ultimately the use cases that are enabled on these unified insights. That’s when you start realizing the ROI from hence getting these games together, the gate is very important so that you have your use cases this year and then against those use cases, you start evaluating the features of the CDP partners that you are evaluating. If there are 10 partners, then you can see which of those partners’ features align to your business use cases. And whenever you see a lot of overlap between your use cases and CDP capabilities of an external partner, that becomes your point of consideration whether each partner would actually fit. And then after that, there are other specific points that you can look at, which is what kind of resolution capabilities they bring in, whether it is deterministic or a longer deterministic it is also probabilistic and so these are IT resolution capabilities. So, what I mean by probabilistic and deterministic is that if Santosh is the customer right now a bring might have several customers named Santosh. But then let’s say if CDP sees that a customer has given a name Santosh and let’s say surname is also given, and along with that, that particular pin code is also given. So that narrows down the entire search, saying that this Santosh is likely to belong from this location and hence this is one of my existing customers. So that becomes a probabilistic part. If a brand has a need for that, then of course, it needs to be there on their checklist while losing the right partner so that the second third is what is the accessibility to data. So, whether the CDP is a closed ecosystem, where brands get to run campaigns, see those insights, but then not get access to the back-end data that CDP is collecting. But let’s say that’s an important use case for your data science team, where you want them to also be customer ML models. So, then you also need access to all of those data points, and that can also be an important factor, which is going to be accessibility to data or the magnitude to which the CDC is able to democratize your own databases. That is another key factor in choosing. And other than that, I think one very important factor is the real-time insights. You don’t want any insights on the CDP to come, which is why these insights have become stale because of the passage of time. You want the insights to come at the right time so that you can act on those insights on sustainability. That is another key factor which the rank can look at is evaluating city departments. And then other than that, I think, of course, cost plays a big role that goes without saying that’s an important factor.
Santosh: Absolutely. Money matters. Fantastic. Fantastic. Thank you. Thank you for giving us this time. And also, congratulations on being listed in the CDP Institute. I know it’s a great recognition to have and I look forward to having many more interesting sessions with you. Perhaps, we could probably do another round on a specific vertical of interest and give our listeners viewers a more thorough drill down into it.