(Editor’s note: Dave Kellogg is CEO of MarkLogic, an information infrastructure software company. He submitted this column to VentureBeat.)

While I often challenge VC conventional wisdom, I actually have a great deal of respect for the difficulty of what VCs do.  Particularly in the early stages, I think it can be extremely hard to see value that might only become apparent months/years later.  For example, by chance, I got to see a YouTube demo when the company was in its infancy – even before Sequoia made its investment – and while I was wowed by the technology (the demo I saw was on a mobile phone), I never would have imagined at that moment that they’d sell the company 18 months later for $1.2 billion.

Some companies did seem obvious to me.  I knew PeopleSoft would work because I saw too many customers building their own HR apps on top of relational databases in the late 1980s.  I knew Salesforce would work because I personally knew the difficulty of working with an IT department more concerned with financial systems and infrastructure than helping business operations.

In a nutshell, I think what early-stage VCs do is very hard.  I know this both from observation and direct experience as an angel investor.

From VentureBeat
Get faster turnaround on creative, more testing, smarter improvements and better results. Learn how to apply agile marketing to your team at VB’s Agile Marketing Roadshow in SF.

My first angel investment was in a company that built a complete linkmap of the Internet in 1999, but used it to automatically detect and re-route broken links (i.e., eliminate dreaded 404 errors).  It died a few years later.  A few years after that I realized how close it was to greatness.  It had the Internet link graph, it just used it to solve the wrong problem:  link correction as opposed to … uh … search relevance (i.e., Google).  Oops.  But this is the nature of early-stage investments.  You can be miles — or only centimeters  – away from greatness, and not even know it.

But I got back on the horse.  My next angel investment was in a company that wanted to build an OLAP-style cube for content, in some ways conceptually similar to MarkLogic, but with very different technology and go-to-market strategy.  The OLAP angle didn’t play out well for them, so the company did a big shift, using its product as an enabling technology for a call center application that eventually was sold to Salesforce.

Many would-be platform companies end up using their platform as an enabler a given application.  In the XML world, Xyleme comes to mind as an example.  The company started out selling an XML database and when they didn’t work for them, they built a learning management system on top and now focus on that.  Frankly, I don’ t know if that’s working any better for them, but it is a good example of the pattern.

My other “investment” (in this case of advisory time and not money) was in a company that did self-service call-center applications and that had the misfortune of focusing on telecom as a vertical in 2001.  Rather than switching strategies when telecom went south, they gutted it out, and were also eventually sold, but not without a few founder gray and/or lost hairs along the way.

So after a brief fling with angel-ish investing in the 2000 timeframe, I stopped for many years.  Only recently have I started up again.

My first “investment” was one of time, not money, in joining the board of analytics-on-big-data vendor, Aster Data in 2009.  I love the Aster founders.  I think they have awesome technology.  I see great things in their future.

I’ve also done one angel-band style-ish investment in a really unique company that’s still stealth so I can’t discuss it.  It’s cool, weird, visionary, and practical.  (And it’s not easy to find companies you can honestly describe with those four words.)

Going forward, I imagine doing a handful more. So if we know each other from a prior (or current) life and if you’re looking for a modicum of angel-style funding, give me a ping.  Some CEOs golf in their spare time; I’d rather spend mine sharing some hard-earned wisdom.

(Although I do sneak in a little fly-fishing from time to timel)

VB's research team is studying web-personalization... Chime in here, and we’ll share the results.