BT

Marc Fleury on what makes open source business models tick

by Floyd Marinescu on Aug 29, 2006 |
Months after the acquisition of JBoss by RedHat, Marc Fleury has been making tours with RedHat investors - explaining the business model and how it works.   Marc blogged some of his thoughts on the JBoss matrix, providing one of the most coherent explanations of how an open source business model is different from a proprietary license revenue business model around. Definitely worth a read for those interested in how open source business models are changing the face of the enterprise software development industry. The key portion:
So what is the business model? Very simply, we transform the way software is developed, distributed and supported...While there is no free lunch and there is still an economic price to this development, the upside is a highly leveraged development model where quality software and limited resources can achieve large-scale success. JBoss reached 1 million downloads when we were just 2 guys and a dog.

...the power of the model rests in the extremely low cost of distribution and sales. We reach millions of folks with free distribution and then monetize this base. It is a very efficient way to acquire customers. The result is that we spend 30 cents for every dollar of maintenance revenue, while the competition, on average, spends $3 for every dollar that ultimately comes in as maintenance. The downside, compared to proprietary software, is that on average we only monetize 3% of our user base for JBoss and roughly 10% for Linux. This low cost of sales we achieve through mass distribution is what makes the model tick. The customer gets to make up his own mind as to whether the software is any good as opposed to having to go through the vendor’s pricey and biased salesforce. This enables the OSS enterprise sales force to be very effective since they mostly are targeting highly pre-qualified potential customers.

The underlying key to the model however is to already have a popular open source product that will produce a user base that can be monetized.  This was also echoed by Larry Augustin commenting on the growth potential of open source business models:
if the product does not have the adoption and usage to generate pull, the model won’t work. The initial temptation is often to fix this problem by creating “push” (i.e. spending money on BMW-driving sales reps to push the product). While push may be an answer, the real answer needs to come from an understanding of why the adoption isn’t there. Why aren’t people downloading, installing and using our product? How can we better understand the needs of the user, and make our software more useful to them so they will want to use. After all, we’re making the software and source code available for free. Price is not a barrier. If people won’t use our software for free, how much good is push really going to do?

This is good advice for companies looking to 'dump' products into open source - don't do it. If the product isn't compelling to begin with, it won't necessarily work as an open source project.  In the Java community, many point to Apache Beehive as such an example, others have suggested that that Glassfish could be such an example.

Hello stranger!

You need to Register an InfoQ account or or login to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Tell us what you think

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

Pre-qualification and evaluation downloads by Rickard Öberg


This low cost of sales we achieve through mass distribution is what makes the model tick. The customer gets to make up his own mind as to whether the software is any good as opposed to having to go through the vendor?s pricey and biased salesforce. This enables the OSS enterprise sales force to be very effective since they mostly are targeting highly pre-qualified potential customers.

How is this different from having an evaluation download of commercial software? The above doesn't seem to be something which clearly separates the OSS model from the commercial license model.

It is also worth mentioning that the OSS model is more appropriate for certain kinds of software than others, and that the distributed development model is more appropriate for certain kinds of software. It's not a one-size-fits-all, but ultimately it is also not the model but the actual individuals and their work which will make the difference. For example, infrastructure is thought to beone of the main domains of OpenSource and yet Tangosol has proven that they are unbeatable when it comes to making a distributed hashmap. Atlassian has similarly shown that the combination of evaluation downloads, a good community, and easy-to-use software can beat the big boys, both in the OpenSource and commercial sphere.

From my reading of what Marc writes, there is nothing there which is really repeatable. There was a lot of tugging on appendages of self, but nothing which really described a model that could be replicated, and nothing that was really unique to a so-called OpenSource "model".


The last statement is one many consider obvious. We transform the way the industry does support. Why? Well, that is the only place where we make ANY money at all, so we better be good at it. End of story. If we are not; well, very simply, we will die.

Here is an interesting enigma that proponents of the so-called OpenSource model does not address, or at least I have not seen it addressed anywhere. If the ONLY place you make money is support, what is your incentive for making the software work? Or work well? Or to make it easy to use? Or to make it so that the customer is independent?

For myself, since we are a commercial license company we have a REAL business incentive to ensure our support is minimal (preferably NONE; customer buys, installs, end of story), that the product is easy to use (based on customer stories and product comparisons we are far far ahead of the game in our market), and that the customer is as independent as they can be (almost all tasks in our shit can be done through point'n'click UI).

The end result is that our software is, generally speaking, an order of magnitude cheaper than similar OpenSource solutions *if you look at Total Cost of Ownership*. I'm not saying this has anything to do with the commercial model, per se, because we certainly have commercial competitors who are producing crap that costs a lot to maintain and support, but at least we are using a business model that is not directly contradictory to the goals we want to achieve.

So, I think it is important that the business model does not, on paper, contradict the goals of giving your customers good value for money. The OpenSource "model" (I'm using quotes as I don't see a repeatable model in what he describes) Marc talks about does not pass this test. Considering Marc's references to the work of Nash - who btw in essence was a psychopath as described and chronologued in "A Beautiful Mind" - this may of course not be your main goal, especially if you have venture capitalists at your doorstep wanting to cash out as much as possible. That is fine, if one is in such a situation, but don't be surprised when customers turn to other vendors after a while.

Re: Pre-qualification and evaluation downloads by Floyd Marinescu

Rickard - thanks for the thoughtful and educational reply.

One point I am curious about is the suggestion that nothing about the process is repeatable. Surely it is - SugarCRM doing a similar process with success, and the ServiceMix/Mule guys seem to be doing the same thing as are a number of others.

While none have had a successful exit/acquisition like JBoss has (yet), the business model itself appears to be repeatable (and if it wasn't, I doubt that VC's would be investing in some of these companies).

Good points about cost of ownership!

Floyd

Re: Pre-qualification and evaluation downloads by Jason Carreira

I'm not sure about Mule, since I haven't used it, but the ActiveMQ/ServiceMix stuff is NOT easy to get up to speed on. There's not that much documentation, mailing list answers are spotty at best, etc. Don't get me wrong, we still chose these to use since overall, for a team with good people who can muddle through and figure out how to use under-documented stuff, it still provides value. If we weren't that sort of team, it would be an interesting TCO comparison for ActiveMQ with service from LogicBlaze vs. a commercial vendor with documentation and service that come with the license fee.

The value of open source tools goes down dramatically, IMO, when the team isn't good enough or have enough time to dig into the source that comes with the OS product to find issues. In that case, what difference does it make if it's open source or not, since you can't fix your own bugs?

Re: Pre-qualification and evaluation downloads by Rickard Öberg

One point I am curious about is the suggestion that nothing about the process is repeatable.

I was talking about the model described by Marc, explicitly, and not some other conceivable model based on OpenSource. Heck, my own model is based on OpenSource, and we're doing pretty good and I definitely think it's repeatable, but it's not what Marc described. See the difference?


Surely it is - SugarCRM doing a similar process with success, and the ServiceMix/Mule guys seem to be doing the same thing as are a number of others.

See Jason's reply, and we are also talking about a relatively small number of example companies here. It is hard to distinguish what makes it a good sign that the model is good, as opposed to being simply good people and good work, possible even in *spite* of a bad model. In other words, instead of looking at the relatively few examples that exist I would like to see something inherent and obvious in the model itself that makes it likely to be repeatable. That is what I am missing.


While none have had a successful exit/acquisition like JBoss has (yet), the business model itself appears to be repeatable (and if it wasn't, I doubt that VC's would be investing in some of these companies).

My experience with VC's is that their view of reality is rarely connected to any kind of actual objective reality. They like to think with their guts, because that's where all the good decisions are made. But your mileage may vary, I guess.


Good points about cost of ownership!

'tis an obvious but often forgotten and/or ignored and/or misunderstood factor.

Business Model: Support and Confidence by William Louth

Is this the business model?

www.jboss.com/index.html?module=bb&op=viewt...
Customer/User: "i hate spending my time with garbage pet projects and i already spent my time with pet hair profiler around 3-4 hours."
Business: "You want that level of support/confidence... go buy a license of a commercial product!"

Re: Business Model: Support and Confidence by Rickard Öberg

Is this the business model?

Customer/User: "i hate spending my time with garbage pet projects and i already spent my time with pet hair profiler around 3-4 hours."
Business: "You want that level of support/confidence... go buy a license of a commercial product!"

If this difference between what they are saying and what they are doing wasn't so sad it would be funny. Now it's just sad.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

6 Discuss

Educational Content

General Feedback
Bugs
Advertising
Editorial
InfoQ.com and all content copyright © 2006-2014 C4Media Inc. InfoQ.com hosted at Contegix, the best ISP we've ever worked with.
Privacy policy
BT