Products

LeadsRx Attribution™

Measure across ALL Channels

5

LeadsRx Privacy Studio™

Establish a privacy-first approach to data

5

LeadsRx Journey™

Optimize conversion paths

Case Studies

Agency Ups Client Perfomance Through Analytics

View All Use Cases

Products

LeadsRx Attribution™

Measure across ALL Channels

5

LeadsRx Privacy Studio™

Establish a privacy-first approach to data

5

LeadsRx Journey™

Optimize conversion paths

Case Studies

Agency Ups Client Perfomance Through Analytics

View All Use Cases

Jan 1, 1970 | Podcasts

Kaz Ohta with Treasure Data and Utilizing a CDP for Marketers

LeadsRx:

n

Hello and welcome back to the Attribution Marketing Podcast, powered by LeadsRx, where we help businesses, brands, and entrepreneurs get the most out of their marketing and advertising spend. On today’s episode, we have Kaz Ohta, who is the CEO of Treasure Data, a leading customer data platform. Kaz really has been a pioneer in the big data space as well as the open source space. And so we are going to tap him for his expertise and knowledge about how to deal with an ever-growing challenge of consumer privacy marketing data, and most importantly, what you can do with all that information to be more successful in your marketing and advertising spend.

n

I think we always wanna start off with kind that career area. We were doing some research before the show, and I think you have a really interesting story that we all would love to hear about.

n

So take it from the top. How did you from you know, your career in Japan versus coming over here to America, learning English, and re-emerging as a CEO of Treasure Data?

nn

Kaz:

n

So, you know, I was born in a rural area of Japan, called Osaka, and uh, I guess, you know, my family has a little pharmacy business there. I was born into that family and then I guess I became a businessman because of the influence of my parents. So when I was 21st, I started my own company. It was a software company in Japan, but I had an opportunity to work with Silicon Valley-based software company that scaled from like, uh, five, five people to 500 people in like three, four years.

n

And my company was growing, okay. It was like five people to 40 people in five. And you know, after looking at the growth and investment in Silicon Valley, I was like, okay, I need to come here. Right? So when I was 25, I came to Silicon Valley and started this company, Treasure Data with two of my co-founders.

n

And my background was in computer science, especially HPC, High-Performance Computing. My professor when I was in university, built the world’s fastest supercomputer at the time, and I was a part of the team to build the file system for the supercomputer. So if you don’t know what is a file system, it’s a data processing system, right? The way the system stores and processes the data.

n

So I learned how to manage and process large amounts of data. So that became the basis of this company’s Treasure Data. Now I couldn’t speak any English when I was 25, but I had nothing to lose.

n

So we scaled from just three people to now more than 700 people across 20 countries now. And it’s because, you know, there’s a lot of growth in the data space. Especially we saw the opportunity of leveraging customer data in the enterprise. So that’s why we’re now providing the solution called CDP (Customer Data Platform), which is one of the fastest growing areas of technology in the marketing space.

nn

LeadsRx:

n

So tell us a little bit more about CDPs. Talk to us about how that may be the next evolution of the marketing tech stack, why enterprise brands specifically need to be paying attention to their CDP, and how to get the most out.

nn

Kaz:

n

Of course. I wanna start from my bad experience with one brand. You know, for example, I’ve been a customer of one cable company for 12 years and I’m living in California. Then one day I got this email saying, Hey, fiber is finally coming to your house. Right. I needed faster internet for work.

n

Also, I’m a gamer, so I wanted to have a low-latency network all the time. So I was pretty excited. So, I called the call center for them. And then, first of all, I needed to wait like 30 minutes and they were asking for my phone number, even though I’m calling from my land line. And they’re asking about my plan, which I’ve been using the fastest tier and the highest tier TV and internet service for 12 years.

n

And they didn’t know about me. And also, of course, they didn’t know about the fiber plan. So I was so disappointed. So I went to the store, the real-world one, and then the same happened, right? So the sales rep is asking for me like the phone number information. They didn’t know about the plan and you know, I need to wait like 30 minutes.

n

So really the purpose of CDP is how can we know customers in a single cohesive budget. And it’s really the basis of consumer expectation, but it’s really hard for the enterprise because the data is siloed everywhere you are running, probably multiple businesses.

n

I saw some companies have like 13 CRMs in the group because they’ve been growing because of a lot of demand, for example, right? And that has been a problem. So CDP really is trying to unify all of the customer data into one place. And then you can use the data across any channel, not just for marketing purposes but for sales rep compliance and privacy reasons, right? So that’s the solution we provide.

nn

LeadsRx:

n

Well said. I think it’s a pretty complicated topic, but think your examples do a pretty good job of highlighting what those use cases are, and I think for each enterprise they’re gonna probably utilize that data may be in a different way.

n

I’m looking at this case study here for Muji who had saved millions in marketing costs. Walk us through how Treasure Data and the CDP really helped them with the marketing.

nn

Kaz:

n

Of course. So a lot of retailers have challenges. First of all, a lot of engagement is now happening in the digital world, right? So according to McKinsey, well, let’s say any business right now, 60% of the customer journey is now happening in the digital world. And this gets tripled after the pandemic. So retailers like Muji, they’re observing a lot of people researching the product.

n

And of course, they have an e-commerce channel so that a customer can buy. But a lot of purchases still happening at the retail level. But the problem is before they come to the store and explore. But then what is happening right now is consumers, especially the younger generation, do all the research online and they come to the store just to.

n

Right. So this creates a complex problem for digital marketing. Okay, what am I optimizing? So the problem first we solved was we consolidated all the customer data, both in digital and also physical channels. And especially the key was their mobile app called Muji Passport. So this allows, first of all, to get the point for the consumer.

n

So that, you know, they’re more motivated to use this app, but really the purpose is to have analytics on top of it, connecting this digital campaign and then physical campaign. We can exactly know who is actually coming to which store, and what type of product they purchased, right? So by using this data, they built a multi-touch omnichannel marketing model so that they can optimize the campaign.

n

Is this campaign actually working? Is this digital campaign actually influencing the purchase in the retail store? Right. So by doing this, they actually greatly improve the marketing ROI through this analytics and implementation of the CDP customer data platform.

n

What when they get involved in the CDP, is it, is it something where you say, okay you have these eight independent tools or silos of data, something’s in SQL, something’s in Google Analytics, Facebook over here, your retail point of sale channel, and then Treasure Data, or a CDP, let’s just say, is able to.

n

Consolidate all that and then deploy its insights and analytics and, and get some value that way. Or is it better stated that you know, the CDP is collecting once it’s implemented new data across all those channels, and then after a certain time period, it’s able to kind of produce its insights and analytics after collecting its own information?

n

Which of those two is more accurate, or maybe it’s a hybrid? Yeah, it’s kind of a hybrid, depending on the data source, for example, the data like CRM, cons, and point of sales, tend to have more historical data available already in the database, right? Yeah. So we can take everything, but for example, the behavioral data such as web behavior, mobile behavior, right, or even in-store behavioral data we can’t go back unless it’s explicitly stored somewhere, right?

n

So we’ll implement mobile tracking, SDK, and then JavaScript. When they deploy these tags and SDKs into those applications, we will be able to start capturing, but we cannot go back usually. Consumer privacy is in focus, not only for enterprise-level, martech, you know, marketing tech stack providers, Google, Facebook, etc., as well as regulations in Europe, California, and all these major markets.

nn

LeadsRx:

n

What do you think is the future of the battle between MarTech providers that want to know exactly in a deterministic way? Bob is Bob, and Bob was here on Tuesday. His interests are ex, he lives at this address, versus what seems to be a growing call to not have that information.

n

And Bob wants to be anonymous. It doesn’t want to tell you. How does Treasure Data get past? What is the future for this type of analysis If cookies go away and everything becomes anonymous?

nn

Kaz:

n

Yep, I’ll answer for the cookie one first.

n

So first of all, just to, for the audience, what we call a third-party cookie is time. So this cookie mechanism is implemented to track the behavior of the browser. Sir Cookie allows a lot of tracking analytics vendors to track across the domain, right? And then apple has already deprecated this, and then Google’s Chrome is supposed to be deprecated in the next 18 months or so.

n

So this prohibits a lot of tracking analytics vendors to track the users across multiple. Right, and the problem with this is you really can’t control it. The behavior and also the conversion. For example, if someone is watching the ads or content in a certain domain and then they ended up purchasing in a certain website, you can’t really correlate if it’s a different domain.

n

So that puts us in a really you know, interesting challenge for a marketer because now that machine learning algorithm cannot learn that converge. The ROI for the advertisement or efficiency of the advertisement ecosystem has calmed down. So that is a huge problem. Now, the trend we’re seeing is what we call first-party data, right?

n

So traditionally, so let’s take an example, like New York Times, right? So they have been relying on cookies to monetize the audience, so anyone can go to their website. And looking at the articles for free. And then in exchange, you will see a lot of ads so that New York Times can monetize it. Now that is becoming a really, really hard or almost becoming impossible model for a lot of media companies.

n

So what is happening right now is a lot of media and brands are trying to have, its. Own customer data by signing in, by forcing users to sign in, right? So if you go to Fox or New York Times, or any website right now you have to basically register first to see the content, right? So by doing this, obviously user has to give some information to the media companies or brands, but at the same time, brands can leverage that data.

n

To have a better advertise you by using that, you know, hash email address or some of the personal information you gave. So that actually gives much higher marketing efficiency, right? So, you know, I’m engaging with a lot of retail companies, G P G A, automotive, healthcare, media, and Entertainment.

n

Right now they are shifting towards direct-to-consumer. First-party data models, capture their own customer data by themselves rather than relying on third-party datasets, such as Google and Facebook because you have to use your first-party data to get the efficiency of your marketing channel.

nn

LeadsRx:

n

What percentage of the people that you talk to or enterprise brands, maybe of a certain level of sophistication, understand? That first-party concept, that they actually have to care about this, collect it on their own, be it custody, you know, take custody of that information, be regulated for having that information.

nn

Kaz:

n

Yep. I would say the buyers of CDP almost hundred percent understand this trend because when third-party cookies are deprecated and when you need to store first-party data, the customer data platform is the place to store it. Right. But when you talk about the broader industry, According to I D C 50% of global 2000 companies like Enterprise Company will adopt CDP towards the end of 2024.

n

So, that’s definitely happening. Maybe more so for the large company that has billions of dollars of marketing dollars because efficiency matters, right? Also what we’re observing. Right now there’s a lot of concern about, you know, ongoing economy slash recession concern.

n

So in our customer base, a lot of marketers are now facing a budget cut or flood next year. And then a lot of their CEOs are asking, okay, is this campaign working? What is what can we do better or make it more efficient? So, A lot of companies are reducing their marketing budgets, but what we’re observing is there are also increasing the investment into data and analytics technologies so that they can get more measurement across all of the marketing campaigns.

n

And that’s why also CDP customer data platform is getting traction because we’re basically tracking every single channel and customer behavior, right? So that will be required to provide the ROI for all of the campaigns we have. If you would tell us just a little bit more specifically about your multi-touch attribution, either approach or the product.

nn

LeadsRx:

n

I see you have these treasure boxes. I like the name, by the way, listed on the website. And one of those is the multi-touch attribution.

nn

Kaz:

n

Of course. So what we first do is collect the real-time and batch data across the channels, right?

n

Whether it’s a web behavior, campaign, data, you know, email ops and ops, clicks and bounce-ins, and everything, right? And then we also take some conversion data coming in, whether it’s POS data or an e-commerce platform. And then by leveraging our ID unification algorithm. We identify, you know, the customer and then create a golden record of customers.

n

We call it Customer 360 view. So, after this process, we exactly know which customer was touching, you know, which channels or campaigns. And then this multi-touch attribution algorithm will create the weight for each campaign or channel. Right. Okay. You know, it’s not like, okay, email actually converts more than others.

n

So typically, let’s say for Treasure Data, we have around 10 to 15 touches before they talk with the salespeople On average. Right. And then it could be anything like webinars or content download or, you know, looking at the email, right? So we will create the wait for each campaign and then you can calculate back, okay, we closed a billion dollars.

n

But then a lot of campaigns influence it. So we will create multiple models around, okay, this campaign could actually be worse, like a 100 K and 200 K, 300 K, 400 k, and then it adds up to a million dollars, right? So that is the core of the multi-touch attribution model. It’s also customizable. You can make a little bit of weight.

n

First touch. Right? What actually generates the first interest or last touch? Okay. What actually makes them convert it? So that really depends on how the company wants to wait and what we learn. It’s different from industries or how someone thinks about the campaign, but the technology at least is configurable to have a different setup.

nn

LeadsRx:

n

Great. How much human intervention goes into your methodology? Is there a human side to interpreting that data and providing it back to clients? Or do you rely exclusively on AI machine learning and what the data tell you?

nn

Kaz:

n

Yeah, so I would say AI and data will actually perform better when you have a lot of data, right?

n

obviously, if you are, let’s say B2B company who have, I don’t know, 200 to 300 customers each paying a million dollar or two, it’s actually really hard to make a pattern based on the AI, and the machine learning algorithms. I see more opportunity of leveraging technology for, you know, B2C company who has hundreds of millions of consumers typically, right?

n

And also at the same time, you know, the content side is more of the art. Right? Okay. What type of you know, content or marketing campaign or wording is actually influencing the buyers? That being said, you know, we are still an evolution of some of the language models like ChatGPT. You already have a lot of content that actually can be generated automatically by those AI and content.

nn

LeadsRx:

n

We also saw a lot of AI can now generate the banners, right? So a lot of CPG directly to consumer companies are leveraging those AI to generate thousands of patterns of videos and images about the specific product. And AI will automatically. What type of butter ads, like the colors or background to deformed?

nn

Kaz:

n

Right, they keep optimizing it. So there is a fine balance between humans and AI. But you know, over time, if you actually have billions of data, I think how to leverage AI and machine learning is more of the productive conversation versus if you don’t have a lot of data and you, you should probably rely on your gut more.

nn

LeadsRx:

n

Yeah, well said. I think if you have a science experiment, to render the answer, you rely on the science if you don’t have statistically relevant data. You can rely a little bit more heavily on human intuition and, and insight. I think shifting gears a little bit here, just curious about some of our audiences in the entrepreneurship space, and I was, I had two questions for you.

n

The first of which is, You know, tell us about your return, as far as I understand, back to treasure data and the investment from SoftBank. If you’re able to just talk us through what led to that investment, what your, plans are for the future and how that sort of deal came together, that would be amazing.

nn

Kaz:

n

Yeah, that is a really interesting question. You know, I was starting this company as a CTO, Chief Technology Officer, and run this company for almost eight years, and we decided to sell the company to the company called arm, which is a cheap company. Everyone is actually using ARM technology. So if you actually look at you know, your mobile phone a hundred percent of the chance it’s actually arm.

n

Right. I think everyone knows Intel for its own CPU, but ARM is actually the dominant CPU IP provider for mobile phones. And when we decided to sell the company arm’s Ambition was to expand its product portfolio, not just from hardware IP, but also the software. But what happened was, Invidia, the GPU company and came to ARM and then said, I actually want to acquire you but I don’t need the software side of the business.

n

So Treasure Data was long story short, carved out again as an independent company. You know, unfortunately, NVIDIA’s acquisition didn’t go through the regulation. That was a different story, but you know, data was acquired and then become independent again. I actually quit the company after 18 months of acquisition.

n

Because I thought, okay, I was doing a good job on marrying two companies together with treasure data, arm in arm. And I was told, okay, let’s split again. So I’m kind of a little bit out at the point. An entrepreneur, right? After that happened, you know, maybe nine months to a year I was looking for the next opportunity as an entrepreneur, of course, looking at, okay, what is my new business?

n

Right? And then, you know, sometime in early last year Masa and kimchi, who are the two top four people at SoftBank came back to us and, hey, you know, we actually want founder. To run the company again. One, there’s a huge opportunity in the data space, right? And then two, they truly believe in the power of founders running the company.

n

Because, you know, MAA himself is a founder and then created more than a hundred billion-plus business by himself. Right. And, you know, I was a little bit hesitant because I thought my journey with treasure data is. But the more I think about it, yes, more data is exploding anyway for the next 10 years, right?

n

And probably more. And also when we looked at the opportunity, right? So on the earth, there are 8 billion people. And 4.5 billion people are internet-connected. Treasure data actually has a fair amount of people’s data inside our system. Combining 450 brands we have, right? So we have an opportunity to influence billions of people, like much more, you know, easier, convenient, and provide a better experience with customer data.

n

But at the same time, We need to be, you know, promoting the ethical use of customer data like we discussed from the privacy perspective, right? So we need to be a custodian of that data. And if you look at the marketing space data and also privacy are the two fastest-growing areas. And then c D P slash treasure data was the intersection of it.

n

So, long story short I saw a huge opportunity and, you know, we’re already a hundred million plus an R&R annual recurring revenue. And then if you look at, you know, global 2000, we probably have 50. So we have 1,950 companies to go. Right? So that’s a huge market opportunity. and for me. Got your work, got your work cut out for you, sir.

n

Yeah. Get out. Yeah. And I was a CTO but next career I wanted to be a CEO just to take on more. Right. So that also matches my ambition as well. My previous co-founder and CEO has now many kids. He’s like, so I cannot do c e o full-time anymore. So he became chairman. That worked out really well.

nn

LeadsRx:

n

Amazing. Well, thank you so much for sharing your story, Kaz, and for being on the episode. We would just want to encourage you, to let our audience know where can they get a hold of treasure data. How can they follow you on Twitter or LinkedIn? Just take an opportunity to, you know, promote yourself and the brand and where you should send folks if they wanna learn more about what your offer.

nn

Kaz:

n

Of course. To learn more about treasure data, c d p Customer Data platform, please visit www.treasuredata.com, www.treasuredata.com. So I’m also personally helping a lot of entrepreneurs and marketers how to get more efficiency, how to build a business, right? I’m on LinkedIn. Just search my name, Kazuki Ohta, and then please say hi and then say, Hey, you actually listen to this podcast. I’m happy to connect.

nn

LeadsRx:

n

Amazing. Thanks again to Kaz Ohta for being on the show. If you wanna learn more about what they have going on, please go to treasuredata.com. This is the Attribution Marketing podcast signing off.

Download Our Attribution Playbook

for Agencies and Enterprise Marketers.