Terry Cunningham’s Adventures: Crystal and Coral8 (Part 6)

SM: You were about to take Seagate Software public in 1998. What happened next?

TC: I made either the dumbest or smartest move ever. In the board meeting I started getting pissed off because people were pushing down the price, playing their games. They were grossly undervaluing the company. If we took our storage business and plugged it into Veritas, the value of the company would be five times greater than the IPO valuation.

We ended up calling Veritas and Business Objects and offered to sell. Veritas was interested, so we did the deal. I went to Veritas as part of the deal and Crystal stayed with Seagate. The deal was announced in the fall of 1998. It was valued at $1.8 billion.

I remember going to sell the deal to all existing Veritas shareholders, and the stock plummeted that day. It was not a pooling, it was a purchase. They were going to have $1.4 billion in goodwill. Seagate was more profitable but Veritas had more top line growth. Then the bubble kicked in. At the peak of the bubble, the Seagate value in the Veritas holdings was worth $22 billion. The problem is that the corresponding market cap was $8 billion. This went on for about a year. Steve Luczo used to tell the Street on every call that if they did not fix it, he would. They then went private. Silver Lake stepped in and bought it. The private equity group acquired the disk drive business out of Veritas as well as Crystal. They transferred Crystal to the private equity team for $1. It was later sold to Business Objects for $1.6 billion.

It was an outstanding deal. Outstanding Veritas’ shares dropped by about 4.5%, Seagate went private with $8 billion in revenue. The private equity group paid $2 billion, of which $500 million was cash. It is the deal that made Silver Lake. I left Veritas at the end of 1999.

SM: What was your next step?

TC: I took some time off and then started looking for something big to do and talking to VCs. I ran into a project at Stanford which was in the idea phase that I thought sounded good. I agreed to fund it and be the CEO. The product is Coral8, which is complex event processing.

It involves flowing data and data in motion. It is an in-memory data processing engine that allows you to query real-time information. It is about processing hundreds of thousands of events per second with sub-millisecond latency. The obvious target is Wall Street, but more importantly it is the core building block of an event-driven architecture. Traditional analytics are periodic batch processes. We have trademarked the term ‘continuous intelligence’ because that is what it is.

SM: Your game plan is to layer your event-driven fabric on top of any business application from manufacturing to Wall Street. It is applicable to whatever their workflow system may be?

TC: Exactly. My last co-op job before college was writing the data management layer for a payroll system. I remember a sales guy coming in to sell a relational database. This is exactly the same. It is a way to process data that is streaming. It is moving, not sitting there waiting to be analyzed. It is adding the dimension of time to a SQL statement. If event A and B but not C in 5 seconds then do some action.

This segment is part 6 in the series : Terry Cunningham’s Adventures: Crystal and Coral8
1 2 3 4 5 6 7