Connected, by mParticle Bonus Episode: Michael Katz and Chee Chew

Chee shares his experience from working at organizations like Amazon, Google, and Microsoft. Topics include the movement of data from afterthought to strategic asset, the evolution of customer data maturity, and the opportunities for customer data in and beyond marketing.

To stay in-the-loop with future episodes and related content, you can subscribe to our monthly newsletter, mPulse, right here.

MK: [00:00:00] I'm Michael Katz. I am the CEO as well as one of the co-founders here at mParticle Today, I'm joined by somebody I get the honor of working with every day. Our chief product officer, Chee Chew.

So before joining us back in may, Chee was the, former chief product officer at Twillio. Prior of that, he was at Amazon as the VP of consumer engagement. Before that he was at Google as the vice president of engineering and site lead for the Google Seattle and Kirkland offices.

So we're going to talk about competitive dynamics, and how competitive dynamics, necessitate teams to create an, execute a sound data strategy in order to compete for consumer mind share.

So before we get fully into it, we'll do just a little bit of table setting. So let's start with the idea of, of competitive advantages. So Peter Thiel famously said, competition is, is for suckers. [00:01:00] Unfortunately for, for most of us building and running companies, we, we do have competition. One of the most, popular frameworks to understand competitive dynamics is Porter's five forces.

I'm sure many people here will be familiar with that, but it was first popularized back in 19 79, 19 80. And obviously a lot has changed since then, but the competitive forces are. Again, there's five of them. It's intensity of rivalry among existing competition, entry of new competitors, threat from substitutes and compliments, bargaining power of buyers, as well as bargaining power of... So as we moved from the industrial age to the information age, we now have to think about how these forces are applied to the digital economy.

And so over the course of the past couple of years, with digital digitization being accelerated competition now comes from really anywhere and everywhere. [00:02:00] And depending on how we view the world, the competition, isn't always what others may think of competition. For example, like Netflix, as long claimed that their competition is asleep.

So it's a battle for mind share it's about a for market share, share of wallet in some cases, time spent awake. So first question I have for, for Chee . So as, as somebody that's worked for some, some of the largest tech companies in the world, Microsoft, Amazon, Google, Probably safe to assume, you know, a thing or two about competitive dynamics.

How should people who are looking to drive growth, think about the role of competition in the digital economy and how does that benefit all of us also as, as consumers.

Chee: Yeah, well, thanks, Mike. It's great to be here. And one of the things that I've definitely have seen an appreciate it from companies I've worked with is competition.[00:03:00]

Is something that helps inspire and drive innovation. But like the core thing is to focus on the value you bring to customers, the advances of of COD infrastructure and the evolution of SAS services that provide like enabling developer, enabling API APIs. It has been incredible. Like nowadays, almost every SAS service you have is starting to about API APIs that allow developers to build on top of it.

And it's lowered the barrier barrier to entry for businesses in virtually every vertical. Traditional legacy services and products now have developers and they use developers as a critical advantage. This means that there's a lot more competition, which means there's a greater importance to innovate on the whole product experience.

And this is what I really am super excited about the growth. Innovation has been around the core product. If you think about it, like. The ride sharing services and traditional taxis, they both get you [00:04:00] from point a to point B, but the convenience experience all around getting the ride from point a to point B that's, what's really made modern righteous services take off and this sort of innovation around the core product, the experience around the product, being the differentiator is a, a huge advance and accelerating and they're in the recent years. And I'm not, I'm so excited about that.

MK: Yeah. It, it's, it's exciting from a business perspective, it's exciting for us as, as consumers. So I think like effectively, what you're saying is that the building blocks are now readily available. The stuff that companies used to have to build and invest millions of dollars of CapEx in are now effectively like they're, they're provided by vendors. Like it's, it's moved to op ex, right. And so if that is the case and I think, I think it is right. Then it really comes to. [00:05:00] Customer experience, right? Like that becomes the primary form of competitive advantage. Is that, is that fair?

Chee: Yeah, I definitely believe that. It's really the experience of the product and the entire interaction with the company around the product. Like consumer is more and more will choose a vendor or a business based upon how they're treated.

Post-sale what's the plasma support, like was the what's the follow-through and, and that's leading business to realize that optimizing that experience is incredibly important and that to make that experience really personal and what I think a lot of the leading Really understood is that at the end of the day, when you think about, about the best customer experience, what one customer prefers may be very different than what the next customer prefers. So if you really want to have [00:06:00] the best possible experience, there's an adaptive experience. That's necessary to understand each customer personally, individually and personalize it for them.

And more and more, it's also not good enough to have a core set of course green buckets. It's not good enough to have like 10, 15, 20 buckets to try to personalize just to that bucket. Oh, this is this class cohort of customers, but how can you get down to a point where you're individualizing it down to the individual?

Like when I was at Amazon, this was one of the primary things that my team focused on the vast majority of the products that you buy and Amazon are available at another retailer. My team focused on making it most personally relevant at every step along the way, the notion was like, how can you make every single pixel tuned for the individual that was the Nirvana every recommendation from the homepage or the product page, [00:07:00] or even their seats.

Like that was what we wanted to do. And at the end of the day, when you do that, the customer experience and we found that to really resonate with customers.

MK: Yeah. So moving, moving beyond this idea of like these rigid templates or personas, right? Like I think we've all probably seen the various marketing brochures or, or whatnot, like the idea of like the soccer mom or the gadget geek.

Right? Like that, that, that to me never really seemed to resonate. And, and, and from a practical standpoint, also, it doesn't really scale all that. That's kind of what you're saying, right?

Chee: Yeah, definitely. And like I remember going through and there's a, there's a classification of like a high-fashion spenders and like people with a lot of money on clothes.

And then indeed there are people who spend a ton of money on clothes. I'm really high clothes. And then there were some of the folks who were like, [00:08:00] spent huge amounts of money on clothes, but they were like really frugal. For socks and underwear, like there wasn't that category of like expensive on this and frugal on this.

Like it had to come down to an individual and she could really tune it that way then you really want,

MK: . Yeah. Human beings are a lot more dynamic than the standard 15 or, or 20 personas that people may have yet lead lead to believe. So I guess like if. Carry this forward CX is, is the moat.

And what that really means is it's about smart, relevant, adaptive experiences. So experiences that aren't one size fits all or one size or maybe 15 sizes, fit all but are tailored to the individuals based on their first party data. What I guess, what, what brands do you think [00:09:00] are doing it right in terms of leveraging their customer data to drive higher quality experiences using personalization.

Chee: Well, certainly I would look first at brands whose core business is centered around serving up content on the businesses who look back on a large Corpus of content and their job is to really find the right one for the customer at that point in time. And when you think about that, you obviously think about companies like Facebook and Tik Tok.

Netflix Spotify, they're pulling from a large inventory of, of content. They're figuring out, like who's asking for it, what they want in that moment, and then trying to get the right content to them. Now, when you first think about that, that's like media companies. When you also think [00:10:00] about it, like modern, e-tailers like Amazon.

We were in a mindset of also serving up content as represented in product pages, but it still was looking at a large Corpus of options and presenting it. so those are the ones that I think have led the way early on, but the opportunity to understand and personalize, I think it goes from everything from that.

Online media content, even too like buying cars real estate pretty much anywhere. I think there's opportunity. If you're looking for an airline flight, like your preferences of what you care about, what you liked, maybe different from mine, almost certainly does hotels, but almost every business has the opportunity, but the ones that are leading the way definitely to learn from, I think are really the content companies.

MK: Yeah, agreed. Agreed. And I think what's fascinating is I think the, the idea of driving personalization based on [00:11:00] your, your preferences or maybe your explicit behaviors. I it's it's, it's nothing. That's really all that new. I think what, what interests me and what I find to be fascinating is when you can start to take into account like mobile, local, Data.

So you open up an app in one location and it's a totally different experience than an app in a, in a different place or, you know, kind of the next generation of personalization.

Chee: Yeah. So that goes to this notion of personalization is about who you are. There's also a lot about what your context is right then and there.

Um, that's also personalization. So location matters. And you'll see that even like Google search, if you go to Google and you search for like a restaurant that opens right now, you'll get restaurant recommendations that are local to you. And so that location is really important. One of the things that we also found at Amazon was when you [00:12:00] were doing a search.

The context of what you had been looking for in the past 30 minutes impacts how you're thinking, like you were searching for a book and now you're searching for like electronics. What happened in the recent past impacts what you're doing right now? So the current context is really, really important as well. ?

MK: Yeah. So context is, has moved beyond. No earlier days of, of kind of mainstream web usage context kind of referred to, to content and what you were consuming now it's about the world around you and potentially even like who you're, who you're with. So there's, it feels like there's this kind of explosion of data it's got, you've gone almost.

Two dimensions to three dimensions. Now that you have like this geospatial component, how I guess, like, how do teams need to start thinking about [00:13:00] tackling the data challenges to get it into a, a good kind of usable state.

Chee: That's really one of the big challenges right now. Like how do you get it into that usable state?

And one of the things that I think about is really thinking forward of what do you want your data infrastructure to look like? And 3, 4, 5 years, what would it look like when your company is twice, three times the size it is today? And the reason I like to think about that and encourage that sort of thinking. Let's start thinking about the marketer.

For example, they're trying to do a personalized campaign and any play time they may say, I wish I had access to X data. I wish I had access to spend data or age information or address or pick pick one. But the most expedient thing at that time might be to go and plumb a line from the data repository.[00:14:00]

And whatever system has it into the marketing system. And you may do that one time and then you do another one, one, and you do another one, one, and then now the customer support team has a similar request. So you plumb one incrementally, one line by one line, and then you end up with this like spaghetti of mess.

So the real question is what should it look. When you have double the number of departments, when you've onboarded more SAS services, as your company grows, how do you want the data to flow? How assessable do you want it? Start from that notion and then work backwards into what is the expedient thing to do today to make sure that you are able to tap into the data that you have.

MK: Yeah, absolutely. And you bring up some great points which I think are kind of near and dear to some of our core beliefs. I think one of the popular misconceptions about customer data platform [00:15:00] space is that it's primarily, depending on who you talk to exclusively a marketing tool. Right. So I guess like, why is that.

Maybe, I think help people better understand, like how should they be thinking about the role of the customer data platform within, within the organization.

Chee: Now look having this customer data and the CDP and all these tools of focus on marketing, like that's kind of the August for step. And, you know, I think that that stunned from the fact that it was the most clearly and immediately measurable.

Impact came in marketing when you paid for a marketing slot and it doesn't convert, you immediately know that you've kind of Tufts money. But when it does convert, then you know that you kind of made money. And so it was like immediately measurable.

You may not see the payoff from a revenue point of view for a while. And so it's a little bit [00:16:00] indirect, went back to marketing, even like before the internet existed, marketers like really understood the value of targeting a personalization, where they were like literally buying bulk mail to send through us mail and they were paying for postage.

They couldn't send mailing to every single person. And then you, as a lot of it was going to be like going straight from the mailbox to the trash. So they knew they had to personalize in order to make their dollars extend. They were hungry for consumer data, decades and decades ago.

And then when the internet came, then the progressive marketers just jumped at the opportunity to get that data, to utilize their and optimize the business. So that's kind of what led. Us to marketing as the first stop, but more and more you're seeing businesses realize that the opportunity is way beyond marketing.

We talked about like content service, [00:17:00] serving business earlier, social media they like live and die on personalization. They live and die on the ability to mine, the customer data, but let's also look at the customer service departments as another example. And legacy companies. The contact center historically is like a pure cost center.

And the name of the game was the squeeze that puppy down to like count the seconds that an agent does on the phone and reduce the number of seconds minutes that people were talking to customers. At Amazon, we realized that it was actually one of the earlier barriers of entry for customers.

To move from physical shopping, where you had a store, you could, drive to, to return a product complaint or anything like that. Like, what happens if you go and you buy it and you get shipped to you and you don't like it, how easy or hard is it to return? I don't know how, if you or folks listening, tell me you have like, [00:18:00] done return on Amazon.

I think that if you have the most likely thing, is that. Once you did it the first time your reaction was, wow. That was actually not so bad. That was pretty easy. And then if you pay attention to what you did, the odds are your willingness to shop went up and your loyalty went up and your spend went up.

Amazon. We built an internal CDP where we validated that we actually quantified the impact over the next year. And then we facilitate as mindset shift for customer service, not to be a pure cost center that you squeezed down, but how does it become a loyalty tool that helps grow the business? This is the power of taking the CDP customer data beyond marketing and integrate it into the rest of the business.

MK: Yeah. So there's, there's a [00:19:00] lot there. It seems like maybe some of that thinking is kind of rooted in maybe legacy approaches where marketing was really, the only means of adjustability and marketing was viewed as a profit center. Whereas other things were viewed as cost centers.

Chee: It is, I will say that I was talking about it as legacy, but the reality is the vast majority of companies.

If you look at all the business of the world, the majority of companies still think that way. I think the opportunity right now for. Folks who are really thinking and starting to understand customer data is to actually be part of the pack. Like some companies are migrating beyond that, but there's a lot of opportunity because the world really is primarily focused on marketing for the customer.

MK: I guess one, one question is. Like, as, as you think about that, the data does need to be managed holistically. Right. But [00:20:00] you have all of these different individual needs of that customer data, different teams wanting to do different things with it, maybe connected to different tools with different specs.

Like how should people think about that? But that process where you have this kind of convergent divergent set of requirements.

Chee: Yeah. So this is definitely changed. I've seen over the past, like 10, 20 years But the problem is interesting, new and different today than it was say 20 years ago. And I think a lot of this comes from just cloud computing and innovation has happened.

If you think about what a consumer data and infrastructure was 20 years ago, there are a lot more like big. Of software, where a company will buy one suite and it will handle everything. And that one suite may have a single Corpus [00:21:00] of customer data that everything pulled from. So there was a true idea of a single source of truth.

But with cloud computing, you have all this innovation where companies are building these amazing specialized. And so you have like a specialized tool for marketing automation. You have one for like ticketing with one, for an aspect of your customer service. You have your sales opportunities. Like you have all these specific tools and each of them have their own data representation of the customer.

And as a partial representative, And so you have, data's kind of scattered everywhere and the UN each department and the company, like they're focused on like succeeding. So they're like gathering their, customers. They're using their tools and they're generating customer data and it's trapped in their silo and it's in their schema.

And so that's, that's the challenge. [00:22:00] And like the, the thing that I think that companies who really succeed at is they look at the whole pie and they look at like, how can we crown all the data? How do we get that data from all the different aspects? It's kind of like, fracking, how do you, how do you pull.

Let's squeeze the juice out of every corner and get into a place where you can really use it. That's what I think is really important, challenging. And, and I think that the important thing is this first start recognize that challenge, recognize that that's going to proliferate is a natural evolution of the world that we're in and then devise a strategy for how you want to corral that.

MK: Is that, I mean, in your view, is that one of the things that make. Customer data. So, so challenging. And maybe as a followup to that, like, what are some of the, some of the common traps that you see teams falling into?

Chee: Oh yeah, that definitely is one of the things that makes [00:23:00] our customer data really challenging.

Because it is scattered throughout the system. Another thing is like each department is collecting just the data they need. Consumers, when you ask them the questions, like they answer quickly, they don't always answer consistently. Sometimes they share accounts. So you might be asking a question, like, what is your age?

And there's actually one person's logged in. And then a couple days later you asked that from. Point and you get a different age and they were actually people behind one account. And so the data is inconsistent. And also with the modern world of people's sensitivity about their privacy that data can be dangerous if you don't treat it well.

So the lots of factors of today's world that make data really challenging and you layer on top of that, our ability to collect. Has just grown immensely that, that you can like a movement track. [00:24:00] You can page track you can page track before someone's logged in. So like they're initially anonymous.

They may convert as you collecting huge volumes of data. our world could not handle this volume of customer data 20 years ago. So the complexity is coming from all around. And then the aggregation of all that I think is what really makes the whole thing daunting. But that's what tools like CDPs are made for.

MK: Yeah. Well, I think one, one of the, one of the things that I know our customers get the most benefit out of is in helping them craft that data strategy and being really deliberate. Just because, just because you can collect certain amount of data doesn't mean you. Right. There's a lot of noise, right?

And there are lots of tools. There's a lot of bad advice out there that says like, it's better to have it and not need it than need it and not have it. [00:25:00] Or it's like, Hey, just, you know, drop this tag and we'll collect everything and then you can kind of parse through it. It, it, it obviously creates a bunch of noise.

It also creates a ton of cost. The thing that we see as like the longer-term impact is like, it doesn't actually force you to be really intentional about creating a data strategy. Right. So in, in making that, that, that leap, how do you think about moving from thinking about customer data as like this afterthought that you can just kind of like

parse through all the, all the noise or rummage through the, through the hay to find the needle transforming it into something that ultimately is more of a strategic asset.

Chee: It is a hard question. And as you said, you can collect a massive amount of data. And if you simply blindly do that, then you're just growing the [00:26:00] haypile from which you're trying to find the substance that you really care about can utilize. So I think one of the most important thing is to really categorize the information that you want. Understand it, normalize it, it and then make it available to your teams in terms of like, how do you do that?

What I would say first of all, is, is important to recognize that it's never too late to start. And it's natural, but that's the companies start off without having a great data strategy. And then you grow and, and then you collect more and at some point you realize that you have a lot of opportunity to mine, but you do have this sort of like disorganized mess of data.

That's kind of natural. Although it's never too late to start the longer you wait, the more silos there will be. So I definitely believe that as soon as you recognize the opportunity, like go at it [00:27:00] conceptualize what you want the data infrastructure to look like when, when your company is two, three times the size of his day, how do you want the data to flow.

One of the things I've seen in businesses that have marketing drive, this is they, they come in and they say, the marketing says, I, I need certain bits of data from departments, a, B, and C. And then because time is crunched and their, our MarTech team really wants to satisfy as quickly as possible.

They take, and they plumbed direct lines from marketing, from the departments, A, B and C to marketing. And that works well enough for now. And then department C comes along and says, oh, I want data from departments. A and B as the most expedient thing is the plum breath lines. And that, as I said earlier, that that creates a bit of a mess, but also creates this other thing that I've seen.

That's been really painful for a lot of companies [00:28:00] is as some point you may be dissatisfied with one of the services and you may want to change that a new innovation has happened. You want to to transform and move to another service. But now that service has many lines of data dependent upon it and dependent upon its existing schema.

And so you're kind of stuck with that service. So you had to make this horrible Sophie's choice of, do you make this big investment to replumb all the lines or do you live with a suboptimal service and that's a terrible choice to have. So I think when most companies really sit down and think about that, they, they understand that having a central system.

More of a hub and spoke model is the way that will really give them the greatest agility, give them great governance and then allow them to move forward into the future and really scale and grow with the [00:29:00] company. So thinking that through thinking what you want things to look like when your company is three times the size this day, and then work backwards from that.

MK: Yeah. And, and I like what you said kind of creating that hub and spoke strategy. Cause it's also about having, having a system that's purpose built to do this. I think one of the things that we obviously see in market is. Almost a, a sunk cost fallacy applied to their marketing tech stack.

It's like, well, I've already invested in an analytics tool or marketing automation tool or an attribution solution or whatever it may be. And if, since the data is already in there, and I know it's going to be tough to get something like this rolled out, maybe I just kind of like route data through that tool, but those tools.

Meant to be like destination systems, not source systems, like that's what they were purpose-built [00:30:00] to do. So the, the misapplication of, of applications as infrastructure may like put a little bit of a band-aid on the problem, but it's not an actual solution.

Chee: Yeah. So it's a little bit nuanced, if you talked to your great engineers who many times a software system starts with that end user application, and then you look at, oh, there's a bunch of infrastructure.

That we would love to take advantage of. Do you build out from the application and sort of tack on and structure, or do you really step back and say, what is the architectural approach? What is the right thing? What is the infrastructure that should sit behind it? Like good engineers will say, look at the end of the day.

You will be happiest if you actually look at it from an infrastructure point of view and see what it takes to [00:31:00] enable the application as supposed to say the application kind of almost does it. How do we just extend that to be infrastructure for us? Like, you'll get a job done for now, but you'll pay an ongoing price and you'll, you'll learn that it was too expensive in the long run.

MK: Yeah. Yeah. The cost of ultimately not being deliberate and intentional by effectively not having a coherent data strategy. So you're constantly in this hyper reactive mode where I need this tool. I need to respond to this thing. I have this pain or problems that I need to solve for versus looking at like, well, how do I build the structural capacity?

How do I start with. The right information, the right data to power my information system, which can ultimately power my communication system. And eventually like my strategy and tactics and whatever tools I use to, [00:32:00] to deliver upon those. I guess like that, that mindset shift. So in, in your view,

what's preventing more companies from getting there sooner or faster.

Chee: You're, you're so busy and you have like these quarter targets or like short term goals. And the question is like, what do you do now? Do you make this trade off of like, I can do the expedient thing now, or I can step back a little bit and look at like, how do I mine and improve the system for a longer term period of time.

This is one of the things that I really learned and appreciate it from how Google and Amazon, both I've worked in terms of saying, look, it's really important to think of. The fundamentals of the business and the fundamentals and the technology side of the business. [00:33:00] And when you make the investment in having great infrastructure it not only solves your problem, but also spawns, huge new opportunities.

The agility of Amazon to like pop up new ideas, new products. I think a part of that has come from the separation of infrastructure that they made from the application load layer and to really value it. Fundamentally that's how AWS came about. And I think that's really, really important, but the thing that often holds people back is in the short term, you look at the unorganized mess of data that you have, and you're concerned that it's just too daunting. It appears a bigger problem than potentially actually it really is. And so you take the most expedient path.

MK: Yeah. And there's, there's some like. Old saying is like, it's, it's not important until it's important or like it's not a problem until it's a problem. Right? Along along those lines, I think that there's a ton of value that can [00:34:00] be created by the customer data platform, as it relates to going out and playing offense.

Right. Creating better customer experience, driving better personalization. Flipside of that is also protecting against privacy. Right? So not sacrificing defense while you go out and play offense. So with the, with the shift in consumer sentiment around privacy, all of the regulation and the changes that have been enforced by apple and well soon, soon to be Google around privacy.

What role should the CDP play in terms of helping teams? Future-proof their business and safeguard themselves from unnecessary risk?

Chee: Let's step back a little bit and say, when I talk to businesses about like, how do they deal with their privacy policy?

And how they enforce it. Unfortunately, there are quite a number of [00:35:00] companies who work by establishing a policy and trying to push that policy out through education and the way that they enforce that policy is by auditing. Like, that's a very weak way of deploying and then enforcing a policy. And that's how, when you have mistakes happen, like there was a privacy violation and one department somewhere like that reflects on entire company, you're only as good as your weakest link, but a lot of times that happens, those mistake happens because like, it was sort of open for human error.

Because the policies were just education for individuals and there wasn't a proactive mechanism to enforce policy. So any great it CIO or CSO will tell you that if you really want and care about these [00:36:00] policies, these are critical for your company. And I think that safeguarding customer data is critical to people's companies.

Then it can't just be education. It actually has to be technology enforced. And in order to do that, you have to have some central management system. That's what, like CIO tools are all about. Like how do you centralize manage the policies, the information policies for the company, and then proliferate that through all the systems, customer data is no different and the role that a good CDP really needs to play.

It's a place where you can establish what the consumer data policies are, , collect that and then enforce that and make it so that people down the line all well, meaning won't accidentally make mistakes and violate policy.

MK: Absolutely. So that, that proliferation and enforcement ultimately [00:37:00] requires deep integration.

Of of those controls rather than I think what you're saying is like, rather than have it be maybe a scan that happens after the, after the fact, right after the violation has occurred, you want to, you want to prevent the issue, not just scan for the next issue.

Chee: Right. Otherwise you're just cleaning up to be in.

Right, right,

MK: right. Yeah. So you, don't similar to, I think one of the earlier points around getting leverage from a tool like a CDP that was purpose-built to integrate customer data into other systems versus maybe taking like an application first approach and then using whatever connectors have been built or, or web books to, to move data around.

That never really gets you completely off the hamster wheel. Right. You have to think about, yeah. The leverage

Chee: it doesn't. And that's from a protection point of view. And then [00:38:00] the thing that I love about great CDP is. It's not only protection, but also creates opportunity. Like, you know, like sometimes tools like are either opportunities or their protection and, you know, like you, you make a hard trade off of like, where do you want to spend the money?

Defensive offense. But customer data has the opportunity for both be dangerous. There's also this massive goldmine. Right?

MK: Yeah. So, You have marketing, you have analytics, you have customer support, you have compliance, you have data engineering, right. But, so where does, I guess, where does the data strategy reside and like who, who should own the data strategy?

Cause like when, when everybody owns it, nobody owns it.

Chee: Yeah, there's a, there are lots of different approaches for this. I seen it owned by CIO. I seen it own by like a chief digital or chief data officer. Sometimes if there's [00:39:00] a like marketing engineering and MarTech team, like they may own it, especially initially when it's driven by marketing.

There are pros and cons for each. I don't know that, like I would I don't know that I would push a big change in that ownership unless like really your, your, the companies is feeling the pain around that. What I would more emphasize is, think about this as any huge asset and risk of the company.

If, if you have another like huge opportunity to asset of the company, you take care of. You make sure that it's not a side hobby, it's not a nice and weekend job for someone, but you actually like to have someone that you trust really own it. I think that's the most important thing. Like recognize it as a huge opportunity opportunity.

Recognize it as a risk. Staff it with someone you trust, to lead it and fund the resources around it to really get the [00:40:00] most out of it. I think, you know, they've asked the most.

MK: Yeah. I mean, we've for the past few years have championed this idea of like data as a team sport, but effectively, like there needs to be a captain of the team.

Yes. Yeah. Makes, makes sense. So we've talked about the importance of, of data quality. We've talked about matters of. Privacy and, and governance. And obviously the need for a system that was purpose-built to, to easily integrate data into the downstream applications. So say organizations get to, to a good state.

They implement a CDP, like, like mParticle, we start streamlining their, their data flows and protecting data quality and making sure that. Violating, any legislation, like what what's next? Like what's what should this be? The setup for?

Chee: Well, [00:41:00] that's the last, that's the super exciting, like extracting the value.

The it's the data science, the analytics on doing, making predictive and prescriptive analytics to understand how to drive, shape the business. That's, that's a huge opportunity perhaps a great topic for a future session. But you know, for now, what I leave you with is, is sort of disclosing the.

The dirty, frustrating secret that a lot of the data scientists of the world has have realized, you know, these data scientists, great data scientists , they get excited and they learn and they get hired for building these great models to take advantage of the data. They're told we have a massive amount of data.

You can really like go wild and build incredible things. And then when they join your company and there isn't this data strategy. Well, they ended up doing is they spend months and months and months like [00:42:00] getting to the data, getting like privacy teams, permission that say they're allowed to use the data in a certain way.

Checking whether their terms of use agreement with the consumer allows these, that data in that way. Negotiating with each department, can I get access to this data? Will you let me have it getting all together? Collecting it reconciling it, fixing a complex cleaning and organizing data literally takes months and months and months.

And then only after you do all of that, then can you actually start building models and that's super expensive to pay data scientists to really be cleaning the data on that's not what you want them to be doing. You want them to really be building and like the what's next is. To innovate and being thoughtful and planful around the data is what accelerates that still.

MK: Yeah, [00:43:00] because especially within the context of a customer data platform I, I think the idea of like garbage in garbage out resonates fairly well. Like I think it's a kind of relatively straightforward concept. But I think what people don't necessarily understand is like when your, when your job is to then connect data to a dozen or couple dozen downstream systems, and then you're also getting data back from those systems.

If the data is bad and then it's being distributed and decisioned on, and then you're getting that return path. Then that derivative data is also bad. So things can spiral out of control really fast if you're not building the necessary foundation for long-term success.

Chee: Yeah. This is, I think one of the dangers of big data, small data, when you're looking at a few things, like it's easier to spot and say, you know, [00:44:00] there's something wrong here.

It doesn't smell. Right. And then, so you can analyze it. You can figure out what it is, debug it. And then it's done the challenge with big data. Is that. So you can sometimes say, oh, well, there's so much data here that like, it's gotta be right. But, but if it, if there was an error, as you said, downstream, and there were like subsequent data that built is built upon it.

Like that data actually could be wrong, even though you have a lot of it, you just have a lot of wrong data. But you can be convinced because of how many data points you have, then you've got to believe it. And so getting it right upfront is so important. .

MK: Couldn't agree more.