Savio, Savio, Savio. How many times must we go over this? Just because IBM can't figure out how to directly monetize open source doesn't mean the rest of us have that failing. Alfresco is 100% open source. Yet we just quadrupled an already healthy 2006 number.

You have stated a truism above (that there must be some differentiation). However, you apparently don't appreciate the fact that differentiation doesn't have to mean proprietary software. It can, but it need not.

Take off the IBM blinders. There are lots of ways to monetize open source without looking like a 20th-century software company. JBoss got to ~$60M by giving away lots of free stuff. Red Hat is topping $500M and shows no signs of slowing. You keep suggesting that the model doesn't work, but the market keeps disproving the thesis.

Google could give away 100% of its software today and still make billions. Red Hat makes hundreds of millions by giving software away. Perhaps you're not groking that service can be proprietary, too, and it need not be a person writing code or answering a support call.

Perhaps, just perhaps, the future belongs to proprietary services, not proprietary software. That's the world I see when I look around. It's lock-in derived from satisfied customers. You know, the exact same lock-in that the vast majority of the world's industries rely on. Proprietary widgets? That's a 20th-century software anomaly that lasted for two or three decades. The open-source model? It's been with us since the world formed.

In other words, look around. The world economy is a service economy, whether that's in putting together cars (Automotive), sending people places (Planes), feeding them (Grocery, Restaurants), or clothing them (Retail). Software was an anomaly. It's not anymore.

Matt, I can't help feel that Savio is suffering from projection...in that since he can't figure out how to monetize OSS...then everyone else must not be able to either.

Also, are you getting as impatient as I am waiting for Savio to get to the point of his last 3 months of blogs on this topic? It's somewhere in his "Emphasis on products" theme of his last few blogs I think. He sees an important distinction between support and product but is having a hard time describing what's so important about the distinction.

So, let me pose this to Savio: What if...I realize this is crazy thinking here...the Subscription was the product...and that the actual software bits were just one moving part of the service? Emphasis on service. ;-)

If that doesn't sound crazy, then please (re)read my blog:
What's in a Subscription? http://connollyshaun.blogspot.com/2007/09/whats-in-subscription.html

For how to grow from one project to many:
Open Source Strategy: Freeing Great Technology
http://connollyshaun.blogspot.com/2007/03/open-source-strategy-freeing-great.html

I know...I know...but will it scale? How about we look at Bernard Golden's thoughts:
Golden: Commercial Open Source - Can It Scale?
http://connollyshaun.blogspot.com/2007/12/golden-commercial-open-source-can-it.html

Finally, I don't get the "Cash is King" prediction. In my experience, the OSS subscription model is a very nice cash flow model. It's a lot more stable and predictable than the license model that can get caught in the "one or two big deals to save the quarter" dilemma.

Happy new year gents (Matt & Shaun) :-)

My views on the distinction between selling support and selling products have nothing to do with what I'm seeing at IBM. I've been reporting how we've been using OSS to grow our total App Server revenue since 2005. Shaun, are you serious that we don't know how to monetize OSS? That is the funniest thing I've read all year. :-)

Look, my point is that there is a huge difference between selling support around something free (i.e. what Alfresco is doing now and what JBoss used to do) and selling a product under the guise of a subscription (RHEL & JBoss today).

I truly believe that the latter is a viable and scalable business model. This is based solely on the fact that it's worked so well for Red Hat and that JBoss tried to stick to a support business model and couldn't scale it beyond ~$30M before RH moved them to the Fedora model. Matt, I know you're aware of other OSS vendors considering some form of gated access to a product as they grow their business.

Matt, while I congratulate you on Alfresco's performance to date, I'll wait to hear from you when you guys are closer to $250M+. I don't know a lot about the ECM market, but I'm going to guess that when Alfresco is at $250M or so, you'll have run out of users willing to pay for support. This doesn't mean you'll have run out of users...just those willing to pay. While you may have paid $20 for the Coldplay record, the average person got it for free. It's human nature, and I don't know why you two feel enterprise buyers will act differently.

Red Hat doesn't sell a service. They sell a product. It's the smart thing to do. Giving away a valuable product such as RHEL or even (gasp!) JBoss and charging for support provides little incentive to purchase from the vendor. Shaun, I'm sure you can look at the JBoss conversion rate prior to the Fedora model vs. now to see what I'm talking about.

To be clear, I am not saying that a support subscription business model sucks. It appears to be the best OSS business model for early stage vendors. But when the vendor grows to ~$250M+, you'll see more and more of them offering gated access to products. Some will follow Red Hat's lead and do so under the guise of a support subscription. Others will follow MySQL's lead and be upfront about selling you a product. Either path works for me.

Btw Matt, I don't know where you get the $60M revenue figure for JBoss. The last I heard from public info, RH was on the hook for an additional $70M if JBoss achieved $30M in revenue (in 2006 I think). However, it was reported that RH didn't have to pay that extra $70M as JBoss did not meet the stretch goals. And since RH has declined to talk publicly about JBoss performance, I'm really curious how you get to $60M in JBoss revenue.

Hi Savio,

Couple of points. IBM surely knows how to menetize OSS...it's include it as Bluewahsed piece-parts within a closed source, proprietary offering has definitely worked for you guys. My point was on the fully open source models that Alfresco, JBoss, Red Hat, etc. use.

Re: Matt's #'s...as part of a publicly traded company I can't shed light on our revenues...but at JBoss, we very actively managed against Bookings, not Revenue. Bookings drove cash flow and since we were aggressively growing, we hired just behind the bookings curve.

Also your "under the guise" phrasing makes it sound like there's a huge difference between the models. Since GPL / LGPL are not designed to control usage of bits (i.e. users can continue to use well after subscription terminates...unlike most proprietary licenses), this is why I continue to state that the product IS the subscription. Also note that our subscription products include full source code...so they are not obfuscated.

We're still in the "Iowa caucuses" phase of measuring 2007 company-specific results in the IT industry but it looks like IBM might have even passed Red Hat as the largest "OSS-related" company worldwide as measured by revenue. If not, it's within spitting distance.

Sun certainly popped up the leader board as well (simply by changing its terms and conditions). Oracle is not far behind all of the companies discussed in the threads above.

This is measuring the individual companies using the same rules that some of the debaters involved here like to use when talking about the OSS ecosystem "growing rapidly" and "beating the proprietary companies" (to become 2%/4%/10% of overall worldwide software spend by 2012).

I can't figure out whether those factoids will help settle this discussion or make the argument worse. But the facts are the facts.

I am on vacation on the beach in Mexico, but wanted to take a minute to add my two pesos to the conversation (especially since it references my comments to Savio's earlier post). I just finished a three margarita lunch, so I hope that you will excuse me for slurring. I am using part of my vacation to read Nassim Nicholas Taleb's, The Black Swan: The Impact of the Highly Improbable. It is an excellent book that I highly recommend. One of the subjects it covers is the fundamental human failure to recognize something totally unexpected, e.g. the "black swan". The fact that software business models can change so dramtically was and is unexpected, and many will fail to recognize or accept the dramatic impact of the change. That this debate is happening is certaintly expected and predictable, but the open source genie is out of the bottle and will not return.

Bill Miller

Hi Dennis,

I've actually lost track of what us debaters have been talking about. :-)

My friendly banter with Savio is around trying to ferret out his key point. He's had a series of blogs recently that dance around the "huge difference between selling support around something free and selling a product under the guise of a subscription".

I apologize for being dense but maybe I can't see it as clearly as Savio does since I lived it the past 3.5 years. Yes....the model naturally evolved over time...that's what happens when the product offerings, patch/update processes increase in breadth and depth in order to serve more and more customers.

The point I keep reiterating is that while the original JBoss model was a subscription around freely downloadable software...there were always a "software bits" part of the subscription since we created and maintained regular patch releases for our subscribing customers that were not posted on SourceForge. Since joining Red Hat, that part of our model has been further refined to include the initial software bits distribution (not just patches and customer-focused update releases).

Re: your point of who the largest "OSS-related" company worldwide as measured by revenue is...that all depends on what revenues you count as OSS-related, of course. But that is not really the angle I've been debating. Big companies (like IBM) typically have big revenues.

Anyhow, my focus at JBoss remains on how I create subscription and service offerings that deliver real customer value. If I can eye-poke BEA or IBM while doing so....all the better. :-)