Wednesday, September 02, 2009

Why Mark Logic Isn't Bigger Than Oracle. Or Is It?

I had a great meeting with a Silicon Valley technology executive this afternoon. He was a very clever chap with all the right credentials: an engineering master's, a top MBA, experience at Google. The works.

He asked us the standard set of questions that tech executives do.
  • Is this a database or search engine? (Yes. As in both.)
  • Does it have transaction consistency? (Yes.)
  • Does it provide real-time search? (Yes.)
  • Does it have a query language? (Yes. XQuery.)
  • Does it scale? (Yes, to 100+ TB today)
  • Does it cluster on cheap hardware? (Yes.)
  • Do you require schema adherence? (No.)
  • Can you handle semi-structured content? (Yes.)
  • Is it native XML? (Yes.)
  • What database does it run on? (Itself. i.e., it is a DBMS.)
  • Do you use Lucene? (No. It has our own built-in search engine.)
  • Is it a database bolted to a search engine? (No. It's both but as an integral hybrid.)
We went to cover common customer use-cases, talking about the generic reasons why customers buy Mark Logic:
  • Messy XML. (Various structures, unknown structures, changing structure.)
  • Big content. (Tens to hundreds of terabytes and nothing else will perform.)
  • DBMS/search-engine integration fatigue. (Tired of paying the costs of development, maintenance, and synchronization.)
  • Semi-structured data. (Such as the wide, sparse table problem.)
  • "Special" search requirements, where "special" could mean any combination of: structured, parametric, real-time, language-specific, or geo-coded search.
Then, suddenly, he caught us off-guard with his next question:
If you actually do all this, then why aren't you bigger than Oracle?
In the meeting, we stumbled: "uh, gosh, well, that's a question we don't hear often." Now, had I brought my A-game to the meeting, here's what I would have said:
On an age-adjusted basis, we are.
For my source, check out this New York Times post, How Long Does It Take to Build a Technology Empire? The post measures various high-tech companies on their historical revenue ramps, specifically addressing the question how long did it take to get to $50M in revenues?

The answer for Oracle is 10 years. In Oracle's year six, according to the graphic -- which is interactive, you should play around with it -- they were a mere $5M in inflation-adjusted revenues. Mark Logic, in year six (from its A-round funding) is many times that and should break $50M, in my estimation and with continued good fortune, in year 7.

Related articles by Zemanta
Posted by Dave Kellogg at 5:42 PM
Labels: ,

1 comments:

Craig said...

It's an interesting, if not fun comparison. But a quick glance at the list reveals the obvious: quick burn companies like Interwoven who rocket past 50 and even 100 MM only to slow -- in many cases plateau -- at levels far short of the giants. In the search and IA space you have the likes of Endeca, Autonomy and FAST (financial shenanigans aside) who can all claim to have exceeded Oracle's growth rate. Heck, perma-growth DW appliance player Netezza is even showing signs of an outright plateau, albeit more likely due to the economy than Teradata or Oracle. Finite demand slowed organic growth, and of those "rocket ships" only Autonomy, via an aggressive acquisition and market expansion strategy (Verity, Zantaz, Interwoven), has been able to sustain such a trajectory. Marathon records are not awarded based on your time at the 5 mile marker. IMHO, a far more interesting and informative study would trace the trajectory of businesses like Oracle from 100MM to 1B and beyond. Not saying Mark Logic is a quick burn candidate, but it will be interesting to see how the company's growth path differs from its search, IA and specialized DBMS cousins over the next 4-5 years. Let's talk after Heartbreak Hill. ;-)

8:07 AM

Post a Comment

Subscribe to: Post Comments (Atom)