News

How Treasure Data captures, analyzes and uses customer data

We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – August 3. Join AI and data leaders for insightful talks and exciting networking opportunities. Learn more about Transform 2022

Just last decade, businesses knew little about their customers. Today, the world is awash with information from the web, the stores and the world at large. Treasure Data is attempting to solve this problem with a tool that tames the chaos, at least a bit, through well-orchestrated curation. 

Treasure Data uses the letters CDP for customer data platform to distinguish its approach from traditional customer relationship management (CRM) products. The goal of satisfying customers is the same, but the software is designed to handle far greater amounts of data from a much wider range of sources. CRM software focuses on following the customer’s interaction with the business while CDP wants to use all available data to tell a better story about the customer’s needs. 

CDP products are appearing from a variety of established firms and startups. Microsoft and Oracle, for example, are focusing on it. Startups like Segment, Klaviyo, Bloomreach and Insider are just a few of the different companies jockeying to help businesses make sense of the relentless data streams and provide options and opportunities for cross-channel marketing. 

Treasure Data has spent the last 10 years building a platform for collecting different marketing data. Cofounders Kazuki “Kaz” Ohta and Hironobu Yoshikawa have built the company to bring in more than $100 million ARR from clients like Subaru and InBev. 

To understand a bit more about Treasure Data’s vision for how companies are collecting, transforming and understanding multiple data sources, VentureBeat talked with Ohta to understand what it’s achieving. 

VentureBeat: Congratulations on building up such a strong company. What is your secret? 

Kazuki Ohta: We’ve been feeling so much tailwind recently. According to McKinsey, 60% of customer interactions for any businesses right now are happening in the digital world. And this percentage is higher for the Gen Z generation of consumers. So all businesses have to understand the customer with data.

VB: That’s a strong force that’s driving the CDP world. Are there more? 

KO: The problem is, 10 years ago, it was a big data error, where you could collect everything about your consumers and customers and use the data wherever you want.  But 150 of 200 countries implemented consumer data privacy laws like the GDPR or the CCPA. So you have to get the consent and then use the data along with the consent of the consumer who also has the right to be forgotten. 

VB: That sounds like a problem. But you framed it like it’s an advantage. 

KO: The new rules mean you have to manage the customer data in a centralized fashion.  So those trends will continue for the next five to 29years as consumer privacy regulations become more strict. 

VB: How else does regulation help? 

KO: There’s now also a tactical wave where, in two years, 75% of web and mobile users become anonymous because of third-party cookie regulations. Facebook lost $200 billion USD worth of enterprise value a couple weeks ago. A lot of companies right now are shifting the strategy to own their own first-party data. That means they’re not relying on third-party audience data. They’re going to want to own their own customer data by themselves and manage the data and use these data themselves for better marketing, customer service and to support any customer touch point.  

VB: Can you give me an example? 

KO: The problem we’re solving is simple.  I’ve been a customer of one Internet and cable provider for 11 years. And every time I call them it takes 20-30 minutes to wait. And then, they don’t know anything about me. They ask for my phone number even though I’m calling from my phone. That’s because they forward the phone to different lines, right? And clearly they don’t know anything about me.

VB: That’s pretty much 90% of the companies that I call.

KO: The problem we’re solving is in the enterprise; there are so many data silos across many systems, like product data, business division CRM and marketing customer support. Enterprises are increasingly using more channels because the consumer is using more devices. There’s a lot of advertisement platforms and geographies and there’s overwhelming fragmentation of customer data. And clearly the existing solutions are not working based on the example I’ve shown. We have one centralized customer data platform so that we can consolidate all of the data from multiple data sources into it. 

VB: My cable company may not end up truly knowing me, but at least the data will be more organized, and I might not need to wait 20-30 minutes to get something done because the customer service reps are always authenticating me and searching for my data. 

KO: Think about buying a car. You go to websites, research a lot, go to dealerships and then try out some cars. Consider Subaru, one of our clients. The problem they had was that they have multiple product lines like Legacy,  Forester and Impressa, and then each product line uses different advertisements, agencies, AVC and even websites. It’s managed by different teams. And then when you go to a dealership, they don’t know who you are, what you’re looking for or when you’re trying to buy, right? 

So what we have done is consolidated all of the data sources into one and created a Customer 360 view. We’re collecting almost 800 million digital touch points every single day in real time. So every day, Subaru prospects and customers hit the website, use the mobile app or go to a dealership. All this data is coming into Treasure Data in real time. 

VB: Once you have the data, then what happens? 

KO: We apply machine learning models to predict who is going to buy which model or what type of accessories they’re looking for. That used to be done by some human. Machine learning is far from perfect, but it provides greater accuracy.  Then, by passing this information to the dealership sales rep, they can spend their time with you more wisely. 

We can compute that you have a 95% chance of buying an Impressa based on all the data you generated on the journey you had with servers from various channels. So the sales rep can better spend time on you.  

The dealership conversion rate increased from 18% to 31%, and then we generated more than a $800 billion plus revenue increase for one region.   

VB: You’ve mentioned compliance and privacy regulations. How does that factor into this? 

KO: You know InBev, the global leader in beer producing Budweiser, Bud Light, Corona and all these different brands. They generate $54 billion USD in revenue annually. The challenge is that there are so many brands growing separately. They had so many data silos across 40 countries and 500 brands. They didn’t have common processes or data. Each country uses different tools. They also relied a lot on the third-party audience data, but now with  third-party cookie data dying, they need to shift to first party data. That means owning the consumer data directly. 

You mentioned the compliance department is asking the marketing department to have a centralized data security compliance. The governance is  complex for something like GDPR and CCPA. With Treasure Data, we’ve consolidated more than a thousand-plus data sources into one tool. 

VB: So, you’re not so much selling data storage as regulatory compliance help. 

KO: Yes, we have a lot of security features such as processing control, third-party audit access control, role data access audit and log-role access control. Those are some of the features we deploy for compliance. 

VB: How do you deal with the patchwork of regulations around the globe?

KO: We are the only ones who have data centers across the globe. Now, in the U.S. and near Japan or Korea. Some brands like LG have substantial business in Korea. Korean consumer data cannot be stored outside of Korea. So they use Korean Treasure Data, Instant Store and Korean consumer data, while leveraging other data and locations to roll out CDP across the globe.

VB: Alcohol is always tightly regulated, but it’s regulated in different ways in different countries. 

KO: What is interesting is that after the pandemic, a couple of countries allowed them to sell alcohol online because no one wants to go to physical stores. So they needed to shift a lot of sales into digital sales and then CDP was the booster of the digital sales channel.

VB: That’s a shift. 

KO: Consumers are finding it useful. If I look at my behavior too, I’m probably using more deliveries rather than going to the store. So that consumer behavior will continue, and we’re going to see more direct-to-consumer approaches are going to rise continuously.

VB: How do you work with CRM platforms? 

KO:  I think CRM is more than an operational transactional store. It’s more on the frontside managing the records and CDP platforms that are more analytical. This means we’re collecting not only the customer properties, and attributes or fields, but we’re also collecting all of the behaviors as well. Every time they hit the website, go to a mobile app or even drive their car or generate IoT product data, we’re gathering it. All of this data comes into our platform and then that’s very different from CRM. 

If you’re using CRM, you’re going to store 100 million records total. But with CDP, we’re capturing 2 million records every single second at the peak time. So it’s a fundamentally different technology. I would say, from the buyer’s perspective, it’s still a little bit confusing, because you wonder, “Do we have to have CRM and CDP at the same time?” Well, yes, but in the next five to 10 years, that gray area will grow for sure.

VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Learn more about membership.

Source: Read Full Article