Building a Big Data Venture: Hiro Yoshikawa, CEO of Treasure Data (Part 3)

Sramana Mitra: You already started thinking about doing something of your own?

Hiro Yoshikawa: He influenced me to start thinking about it. I have to say that until I met him, I didn’t really have any clear idea about starting my own company. He was the influencer. We had joint trips together. When I took him to Japan, I was introduced to the community leader of Hadoop. His name is Kazuki Ohta. Everyone calls him Kaz.

He ended up becoming my co-founder. Kaz and I have a very good chemistry even though both of us are in the very same space. He was more like a community leader. My background was more in business. In the last three years, I did a lot of venture investments. I started spending a lot of time with him and learned a lot about real-world Hadoop use cases. Kaz was involved in many early stage Hadoop implementation projects. He was clearly very sharp and entrepreneurial. He started coming to my house in Silicon Valley.

When we spent the summer of 2010 at my place in Cupertino, I casually shared my idea of starting a company with him in data space. It was the beginning of Treasure Data. We needed 15 months to get the company going. We started operations in December of the following year after finishing seed financing led by a famous venture capitalist called Bill Tai.

Sramana Mitra: His brother is one of my closest friends actually.

Hiro Yoshikawa: Oh, really. Gus?

Sramana Mitra: Yes.

Hiro Yoshikawa: He’s super nice to me and the company too.

Sramana Mitra: How did you get connected with Bill and why did Bill choose to invest in your company?

Hiro Yoshikawa: One of my mentors in Tokyo is Morio Kurosaki. He’s also running a fund based in Tokyo. He has been long-time mentor. When I shared the startup idea with him, the first thing he did was to introduce Bill to me. They made a joint investment together several times before and shared some success.

Sramana Mitra: Do you remember the idea that you went to Morio and Bill with?

Hiro Yoshikawa: The original idea that I presented to Morio and Bill is not much different from what we are doing right now.

Sramana Mitra: That makes life a lot simpler.

Hiro Yoshikawa: We tried a lot of tweaks but the core vision remained the same.

Sramana Mitra: What was the idea?

Hiro Yoshikawa: Looking at the history of computer, we really believed that computer is the history of abstractions. For example, software was originally just a subset of the mainframe computers always custom-made for each mainframe user. IBM came in and started packaging software to eliminate the need for the custom solution. In networking, some Silicon Valley companies had to design and customize your network from hardware, router, and the TCP/IP stack. Cisco killed the need for this work by standardizing the stack.

Likewise, one of the things that Bill and I have observed was that there were lots of pain points in the market. The biggest pain was the big shortage of data engineering talent while the need for data analytics is growing very quickly, and the modern data structure just got complicated. You cannot run the business just by data infrastructure. You really have to focus on creating values by running analytics. This is the source of business values. We don’t get values directly from data infrastructure investment.

Like Cisco, we strongly believe that we have to abstract the very complicated hard things in the Big Data infrastructure to help our customers or users to focus 100% on driving the real data-driven businesses without thinking much about data infrastructure stacks. This was the original idea. In a nutshell, we wanted to build a very advanced yet fully abstracted, easy-to-use data pipeline management service on top of the cloud. This was the original idea. This hasn’t been changed at all since we started.

This segment is part 3 in the series : Building a Big Data Venture: Hiro Yoshikawa, CEO of Treasure Data
1 2 3 4 5 6