When you talk with a person who has worked for successful companies and spent time as a venture capitalist sitting on multiple boards, you tend to actively listen to the stories and glean what insight you can. When I spoke with Bruce Armstrong, CEO of Kickfire, it was such a conversation.
Bruce is looking to take Kickfire and disrupt the data warehouse market. Their product, as Bruce put it, helps bridge MySQL (which doesn’t understand data warehousing) with data warehousing (which doesn’t understand open source). He is now in the midst of a David and Goliath battle with Teradata — his former employer.
Question: Hi Bruce. Care to describe this battle with Teradata?
Answer: Actually, Tom, we believe the battle is less with Teradata’s existing high-end data warehouse business and more for the vast, under-served data warehouse “mass” market out there. As such, we don’t plan on taking the battle directly to Teradata or anyone else, but rather create a new market by providing affordable, easy-to-deploy data warehouse and data mart appliances for the mass market. The mass market for data warehousing includes rapidly growing small medium businesses as well as departments in larger organizations who simply can’t afford either their own Teradata machine or in a lot of cases even the charge-backs from central IT to time share on a Teradata machine. So, we hope to be complementary to Teradata, but realize there may be some skirmishes for budget dollars in the market today!
Question: You started as employee #16 at Teradata who boasts Walmart as a customer. Was that your start that brought you to running Kickfire today?
Answer: Yes, I was at Teradata for 15 years, Tom. I worked my way up from programmer to President after we went public and were acquired by NCR/ATT. I spent almost 10 years in the field working with customers – including Walmart – so I developed a pretty good idea of what they are looking for at the high end. After Teradata, I was GM of the Server Group at Sybase where one of the products I launched was Sybase IQ, the first column-store database on the market. Later, I was EVP of Sales Marketing at Broadbase – the second column-store database on the market and a pioneer of data marts. Since then, I have been involved in doing due diligence for a lot of the data warehouse startups, including Netezza, Greenplum, and Vertica.
Question: What did you learn from Teradata back then that you are applying to Kickfire today?
Answer: While I learned many things at Teradata that are applicable to Kickfire, the key concept for me was just how different the data warehousing high end is from the mass market. At the high end, once you’re able to break into the market with a new approach, it quickly becomes less about technology and more about services and support. Large customers just simply demand more hand holding. In the mass market, it’s just the opposite. Customers in the mass market expect the product to be highly packaged – easy to buy, install, and manage. They simply can’t afford lots of services and support, so the product needs to deliver high performance in a plug-and-play way.
Question: You have years of history in the data warehouse industry. Care to explain what makes you think the data warehouse industry needs changing?
Answer: While I believe every sector of the data warehouse market is looking for more performance, faster deployments, and lower cost, our goal at Kickfire is not necessarily to change the data warehouse industry but rather to enable more people in the market to take advantage of the benefits that come from data warehousing. Again, we’re going after an under-served market with rapidly growing small and medium sized businesses and departments in enterprises that simply can’t afford Teradata or Netezza. The change that’s required to enable the mass market is to get high-end performance out of a mass-market machine. Kickfire achieves this through our patented parallel-processing SQL chip. In a single chip, we are able to get the performance of dozens of general purpose CPUs. So, in the same way that NVIDIA has radically changed the dynamics of the graphics industry by encoding the graphics processing language in silicon, Kickfire has the same opportunity in the data warehouse industry.
Question: There have been recent acquisitions. Sun purchased MySQL. Then Oracle purchased Sun. Should the MySQL community be concerned?
Answer: We do not think so. We believe that with Solaris, Java, and MySQL, Oracle now has the kind of assets they will need to truly make Microsoft worry. With MySQL, Oracle will finally be able to undercut Microsoft as well as to dominate the high end in databases. As such, we think Oracle is going to continue to invest in MySQL. In fact, Oracle has already proven to be a good steward for InnoDB, the most popular MySQL storage engine they bought several years ago.
Question: About Kickfire, how does this chip technology bridge MySQL with data warehousing?
Answer: MySQL has become the de facto standard for online businesses by virtue of its open source business model and the fact that it has increasingly become production-ready for transaction processing applications. For data warehousing applications, though, MySQL remains very primitive. As such, customers struggle with database volumes as small as 50GB when it comes to reporting and analytics. The primary technical issue with MySQL regarding data warehousing is that MySQL does not have any query parallelization capabilities. This is where Kickfire comes in: using MySQL’s pluggable storage engine API, Kickfire takes over a query from MySQL and provides parallel-processing with the SQL chip allowing queries to run 10x – 1000x faster.
Question: Are you getting any response from the MySQL or data warehouse communities? Are they seeing potential benefit?
Answer: Yes! We have a large and rapidly growing pipeline of opportunities in the MySQL community. One of our early customers, Mamasource, has been using MySQL for years to drive their online community website. However, when it came to analyzing the clickstream data about their community in order to improve the user experience and increase ad revenue, MySQL hit the wall at less than 30GB of data. With Kickfire, Mamasource is now able to run their queries on average 20x faster and up to 600x faster for complex queries. More importantly, Mamasource can now scale their data warehouse to over 300GB, which allows them to increase from just one month of data to a full year.
Question: It’s hard to ignore Microsoft. How does Microsoft with SQL Server fit in the Kickfire world view?
Answer: Microsoft and Oracle are the predominant databases in the mass market today. As MySQL continues to penetrate the market (at a rate of 70,000 downloads a day, by the way), Kickfire will be brought into more and more Microsoft and Oracle shops. As such, one could imagine us building more specific features to co-exist with those databases in addition to MySQL.
Question: You spent time as a Venture Capitalist. We won’t hold that against you but what are your take-aways from being a VC?
Answer: My time as a venture capitalist was really more about board work than investing. As an operator, the VC firm I was with was interested in my ability to help identify interesting spaces, provide input on business plans and teams, and then work with portfolio companies from a board perspective in order to help generate value for shareholders. A lot of what I did was to dive deep into each function (sales, marketing, development, etc) of the portfolio companies and provide input to the boards (including the CEO’s) on where things could improve. Having served on 20+ boards, I got a pretty good idea of what works and what doesn’t work in start-ups and rapidly growing companies. The lessons I learned were really quite simple:
- make sure you have a big market opportunity;
- understand very clearly who your customer is and why they should buy from you;
- build a very high quality product;
- regularly assess your market position and look for ways to change the game in your favor;
- don’t be afraid to tell the truth about the business – there’s always more than one way to create a market leader.
Question: It’s been a pleasure talking with you Bruce. Do you have any additional links or information about Kickfire you want to share?
Answer: Thank you – the pleasure’s all mine. Please come visit us at www.kickfire.com!