Today at OSBC, the results of the North Bridge Venture Partners’ annual “Future of Open Source” survey were released. Try as I may, I could not find out who the respondents were (i.e. developers, CIOs, the Asay family, FSJ, etc) or what the sample size was.

You can see a summary of key findings at the bottom of this page. Here are two findings not included in the key findings list.

Head on over to chart 9/16 to the question asking “Who will command the majority of commercial OS software revenue (non consulting) in 2012?” Yes, 15% of respondents went with pure play vendors like Red Hat. However, over 2x that figure went with platform vendors like IBM, Oracle, SAP (?) and Sun.

Even more interesting, chart 11/16 asked: “Can a startup software vendor realistically enter the (enterprise) market with a product/service that is NOT open source?” Nearly 80% of respondents disagreed with the statement. Now, if you’re a VC, and you see the type of money that closed-source vendors can generate vs. OSS counterparts, (i.e. vmware vs. Xensource) are you going to think twice about the “should this be open sourced or not” decision?

Another tidbit that David Skok (JBoss VC) gave at OSBC was that the JBoss support renewal rate was 85% (likely at the time that JBoss was sold to Red Hat).

It seems strange that a customer would buy support in year 1 and then decide not to renew the support agreement in year 2. Remember, 15% isn’t chump change. An 85% renewal rate means that you have to “grow” 15% just to stay flat with your previous year’s # of customers, or potentially, revenue. In most software markets, 15% is about 1.5x or more of the market growth rate.

Why didn’t the 15% renew?

1] The OSS product is no longer being used, in favour of a different (OSS?) product
2] The application running on the OSS product is no longer required
3] The level of support that a paid subscription/license provides didn’t meet the customer need (either because of under utilization of support or under-delivery of the support experience)
4] Something else?

You can’t do much about #1 or #2, although you’d hope that growing use of OSS, and in particular, your OSS product, would ensure a near 100% renewal rate with customers you already had.

But #3 appears to be a much larger concern. What happens when 15% of your current paying customers decide they can use your OSS product without paying you a dollar. Worse still, these are users you convinced to buy support/license from the mass of non-paying users. Customers surely realize that their support/license payments enable the OSS vendor to continue developing the product in question. Sure, you get some free development from the community, but 95%+ is still done by the vendor’s employees. What happens when more and more customers pass the “pay for continued development” buck and simply become users???

Traditional software renewals rates aren’t 100%. But you’d expect higher than 85% from OSS, since conventional wisdom tells us OSS tracks closer to customer needs and does away with the ‘pitfalls’ of the traditional software business model.

Finally catching up on some reading…I came across this post at ZDNet that discusses an upgrade of Red Hat’s stock by Credit Suisse.

The analyst (Jason Maynard) seems to be on the same point that David Skok (of JBoss VC fame) made at OSBC. How can JBoss/Red Hat monetize a larger percentage of their user base? Currently only 3% of JBoss customers (or downloads? How would JBoss have a count of non-paying customers) actually pay for support.

I quickly thought about Marten‘s customer groupings that he summarized at OSBC.

[1] Those willing to spend time to save money
[2] Those willing to spend money to save time

I knew that I liked his two groups for a reason. I totally forgot that I had come up with 3 groups of OSS users a little while back. My groupings were, more or less, based on willingness to accept risk.

{1} Cost Contentious: Unlikely to ever pay
{2} Scope Contentious: Pay based on app criticality & perceived need for support
{3} Risk Contentious: Likely to always pay (i.e. CYA)

Both views make sense, but willingness to trade off Time & Money doesn’t happen in a vacuum. That’s why I believe willingness to accept (varying levels of) risk is important when trying to categorize OSS users.

On the surface, there is little change to the Time vs. Money equation as a result of JBoss adopting the Fedora model. A customer that wants to spend time to save money can continue to use JBoss “Community Edition”. As a result, one could refute Mr. Maynard’s claims that JBoss adopting the Fedora model will drive more revenue.

This is where risk comes into play. Keep in mind that the JBoss “Community Edition” will include the latest features, some of which won’t make it to the JBoss “RHEL” version or may be dropped from future JBoss “Community Edition” versions. JBoss is clear that backwards compatibility isn’t guaranteed with JBoss “Community Edition”.

If you’re a customer that wants to spend time to save money, and you’re somewhat risk averse, than you don’t like using a product with features that may disappear in the future. You have the option of buying JBoss “RHEL” and getting backwards compatibility. You can also look elsewhere for an application server solution.

Taking Risk into consideration is likely the #1 reason that an OSS user will turn into an OSS paying customer.

 <Updated 2007-05-30> “Risk aversion” shouldn’t be considered a dirty secret of OSS.  I can’t think of an analogy in the traditional software market. But you don’t get to choose whether to pay for traditional software or not :-) </updated>

As a first-time attendee of OSBC this past week, I really didn’t know what to expect. In a word, I’d say I was impressed with the conference.

I’ve never been surrounded by so many lawyers, VCs or startup CEOs for such a concentrated period of time. Most of the breakout sessions were quite interesting (regardless of whether I agreed with the speaker’s pontifications on OSS).

Some highlights (i.e. things that I agreed with):

MySQL‘s CEO Marten Mickos claiming that there is no OSS business model. That we are all participants in the software market; and we use OSS methodologies to develop, distribute and profit from software. Marten also distinguished between two customer types:

[1] Those willing to spend time to save money
[2] Those willing to spend money to save time

He discussed how customers seldom move from group 1 to 2 or vice versa. A very interesting point for folks who hope to convert the 12mil downlowds into paying customers.

The two customer keynotes on day 2, Rob Curley from the Washington Post and Lee Thompson from E*Trade spoke a lot about their use of OSS to drive business results. Rob showed us some really cool news products his team was able to develop using various OSS products. The crowed oohed, ahhed and laughed appropriately. I asked a question that I pretty much knew the answer to:

“Rob, can you talk about how much your team spends on OSS licenses or support contracts?”

This seemed like a hugely appropriate question for a conference attended by hundreds of folks hoping to make a living (and them some!) from OSS products and companies. Rob’s answer was:

“Well, we don’t spend anything. We don’t buy support. It’s likely not the answer you wanted to hear”.

I’m not sure how many in the audience heard his answer or took a second to process what it means for someone hoping to cash in on OSS. I asked Lee the same question in person as there wasn’t time for questions during the keynote. His answer was the same as Rob’s “we don’t pay for anything from OSS, no licenses, no support. We use the OSS project forums for support.”

Anyone else find it funny/strange/unsettling that the customers chosen to speak to a group of VCs interested in OSS and OSS CEOs aren’t spending a single dollar on OSS?

I don’t mean to imply that we should extrapolate from these two customers to “there is no money to be made in OSS”. What we’re seeing here is what Marten described in his two customer types. I truly wonder how many attendees really understood the impact of the two customer types on their aspirations to strike it rich with OSS.

I had a great discussion with Michael Cote from Redmonk about Pragmatic Open Source….Cote’s term for some thoughts that have been popping in and out of my head lately. The discussion deserves a post on its own. For now, I’m just giving Cote props for the term. More later.

Off to the airport!