Part IV of IV: The Battle of the Business Models


Put all of the announcements covered in my first three entries together and you start getting the picture. The current tectonic shift in the market is about business models and the disruptive impact of OSS. On one hand, you have Open Source and Open Standards and on the other you have Closed Source and Closed Standards. But, more importantly, not everyone can play on both sides. The companies that are trying to fight the OSS game, like IBM, are "traditional" software companies with "traditional" software cost structures. They have traditionally high cost of sales, higher cost of development, and high cost of marketing.


When corporate JBoss was two people and a dog, we had reached our first million downloads. It was also already a formidable community, vast, diverse, largely anonymous, but pervasive. Today we are 150 people at JBoss, Inc. and JBoss Application Server (AS) has become entrenched at 40% market share, with BEA and IBM each at 20%. With JBoss Enterprise Middleware Suite (JEMS), which includes JBoss AS, Apache Tomcat, Hibernate, JBoss Portal, JBoss jBPM, JBoss Cache, JBoss Eclipse IDE, with JBoss Rules and JBoss Transactions coming online in Q1 2006, we are accelerating the pace of innovation and the penetration. We are not just winning customers; some would say we are gobbling them up.


How is this possible? While our competition is forced, by their very nature, to fight a war of attrition, we engage in maneuver warfare. Think Vietnam. When Ho Chi Minh recruited General Vo Nguyen Giap to lead the Vietnamese forces against the French and, later, the Americans, he told the story of "The Elephant and The Tiger." The metaphor describes how smaller military forces can defeat larger ones, if the proper strategy is employed. If the tiger ever stands still, the elephant will crush him with his mighty tusks. But, the tiger will not stand still. The tiger will leap on the back of the elephant, tearing huge chunks from the elephant’s side, and then leap back into the dark jungle, as the elephant slowly bleeds to death. The days when my competitors can “JBoss doesn’t have this” or “JBoss doesn’t have that” are numbered: JBoss Rules and JBoss Transactions are the latest chunks we are ripping out of those arguments.


How can a small structure of scrappy geeks challenge elephants like IBM? I believe it is because we were born on the Internet, as a community and as a company. Our cost structure, our organization, our business model is build around the identity of a software vendor that leverages the impact of the Internet. We use it for development using the traditional models of OSS development; we use it to achiever lower costs of sales and marketing; we use it for distribution, reaching mass market at very low cost; we use it for support, offering free forums and for-pay programs to support hundreds of thousands of developers around the world. Welcome to the OSS jungle. We can survive, nay thrive on lower revenues, and that is a crazy new world for our competition.


Our organization reflects our roots. Can any company make the transition to such a business model? Well, not really. The elephants would have to shrink significantly, since our revenues are lower, to be able to mimic our cost structures and compete on a equal footing on the value we deliver. The bottom line is that the Internet has spawned companies like ours and MySQL that deliver value comparable to the traditional players at a much much lower cost. This fight is not even fair.


The move to subscriptions is scary in an on itself. Subscripations are great revenues that are renewable and predictable. It is not inherent to open source. CA, Salesforce, JBoss, Red Hat are all in this mode of sale today. Can companies move to a subscription based business model overnight like SUNW just did?.


If Sun could do such a bold announcement the other day, it is because they DON"T REALLY MAKE MONEY AT SOFTWARE. The softare revenue represented around 1% of their revenue (120M on 12BN sales). So from a Wall Street standpoint it was a wash, it was irrelevant for the money guys--which is why SUNW's announcement generally drew yawns. In fact, Sun was pitching that they intend to increase their revenues BECAUSE of that shift. That is a luxury that Sun has because, ironically, they sucked pretty bad at selling proprietary software.


However, this is not the case of our proprietary competition: BEA and IBM. IBM has $2Bn running on WebSphere and BEA SOLELY depends on its Tuxedo/WebLogic line. For them to transition to subscription and OSS is a systemic shock that might literally kill them. It took CA I don't know how many years to make a smooth transition to a pure subscription business model and they did so by gorging on acquisitions of installed bases in order to increase their base of renewable business. I believe BEA may follow the same tactic of subscription sans OSS, which is possible a la CA. Increase their base to 80% maintenance renewability, "restructure" the last 20% of your work force and, voila, you have achieved the transition.


There are reasons for why the transition is scary. First, a simple accounting rule. If you sell $1M of licenses you usually get to recognize EVERYTHING UPFRONT as revenue. In other words you show $1M revenue in the first quarter you sell. IF you move to subscription you recognize 1/4 every quarter. In other words, you show Wall Street a drastic drop to $250M revenue instead the usual $1M when you do the transition? From an earnings standpoint you drop 75% on that cusp. Obviously it all washes out in the long term, but the cusp is the scary moment for Wall Street, with its short-term focus. Again this was irrelevant for SUNW since licenses represented less then 1% of their revenue. Novell, as another example, did the transition easily to Suse. They claim since they were not selling any licenses anymore, so it was all maintenance anyway.


Add to that the fact that you got to cut your prices in OSS in order to compete with the likes of RHAT, JBoss and MySQL and you get a double whammy effect. Lower revenues right there. And trust me these guys just don't have the corporate culture or structure to handle such shocks: THEY WERE ORGANIZATIONS DESIGNED IN ANOTHER ERA, IN ANOTHER TIME, WITH DIFFERENT BUSINESS ASSUMPTIONS AND DRASTICALLY HIGHER SUPPORT STRUCTURES.


Very specifically, whereas JBoss spends the traditional software P&L on sales and marketing, the orders of magnitude efficiency brought to both structures by the Internet, reduce our sales and marketing costs 10x compared to our proprietary competition. Bottom line we can support SM without a for-pay license, as simple and scary as that (and by DEFINITION of our model BTW). I hope the math is trivially clear. What scares our proprietary competition is not so much the end game of OSS and subscription as Schwartz called out the other day, but RATHER THE SPEED AT WHICH THEY MUST ADAPT THEIR ORGANIZATIONS TO THIS NEW BUSINESS MODEL AND THE FACT THAT THEY ULTIMATELY HAVE TO SHRINK. If they miss the transition, again it could a shock to their systems. This, by the way, is why you should never believe the intentions of IBM and BEA when it comes to OSS middleware and database. It is all about making sure the transition happens on their timeline, that things get commoditized as slowly as possible, so they can internally manage their organizations. I mean... Websphere has 10,000 developers (!!). We on ALL of JEMS have 50. People used to throw this fact at our faces as the proof we would never succeed. It is actually their Achilles heel, now that we are fighting a jungle war on our turf.


Why cannot the elephant fight like the tiger? Assume IBM or BEA resolved the internal political battles over open-sourcing their proprietary technology. Make the even less likely assumption that they could make ditching hundreds of millions of dollars of revenue for a smaller eventual outcome palatable to Wall Street and their shareholders who sweat them quarter by quarter. Then the technical and legal reality is that most of the larger proprietary implementations, especially in spaces like middleware, are neither legally nor practically consumable in open source. So they have to start over, and fight in a new space against an entrenched competitor that is faster and more agile than they are.


I think those cost structures and the inertia built in them (you can't restructure that fast) are the biggest weaknesses of the elephants as we fight turf wars of accelerating timelines of OSS'ing. Between Sun and JBoss we have OSS'ed overnight TWO key components of the high-end technology the elephant is feeding on now, and which they were counting on until the year 2030. Integration and distributed transactions are now out in the open. That timeline acceleration is what scares them to death. How fast we commoditize dictates how fast they must react and adapt their structures to the new realities of the new market. No one, not even us, would have predicted a year ago that in December 2005, both integration and distributed transaction management would be commoditized. In, the space of week, Sun and JBoss just created a nightmare for the elephant.


By publicizing a pro-open source stance when what they wanted was to push Linux, IBM grabbed the tiger by the tail. Open Source has been unleashed, thanks in part to their efforts promoting Linux. Meanwhile, they never intended for their own business to come under pressure. They only meant to hurt MSFT, not themselves...Now, at a time when Sun and JBoss are going all open, BEA and IBM are going closed. They are retreating into their shell to closed source and closed standards. SCA reflects the bare mathematics of their financial structures. They can't afford to commoditize SOA, not that fast. Integration and transactions overnight, they just can't. So, instead, they create an exclusive club around a closed standard and pay lip-service to open source. All the while, they make sure the implementation will remain low-end while they continue milking the high-end. It is a desperate attempt to get control of the timeline and protect the high end. IBM’s attempted play with Geronimo is transparent to the market, why adopt backwards technology that lacks all the goodness, when the stack has already been open sourced? They will do everything they can to slow down the commoditization curve; meanwhile, we will continue accelerating it. The new reality of the market is that CUSTOMERS adopt the timeline as fast as they can. Customer demand will settle the score.


In the meantime, Professional Open Source companies must act like the tiger in order to survive—attentive, observing, fast, stealthy, able to exploit the element of surprise, striking everywhere, able to retreat into nowhere, agile, adaptive, capable of swiftly reacting and re-calculating their moves. The traditional IT superpowers lumber into the OS jungle, where we were born and where we live and breathe and think they can take over our territory because they have more employees, more revenues, bigger balance sheets and a bigger name. While they get mired down in fighting internal political battles and tough external selling jobs to Wall Street, the real OS players will continue to innovate. The entrenched OS player has to fail massively in its delivery on quality of product and quality of support to lose its first-mover, market-leader advantage. In the jungle, speed and agility and knowing your territory win over size. OS does not stand still.


The first time we encountered this “Anything an OS player can do, a proprietary vendor can copy and do better” scenario was when HP started giving away Bluestone for free. It was a defining moment for us. We wondered if traditional logic would hold and the larger brand and the larger balance sheet would win over the established OS player, once cost no longer became an issue. Very simply we wondered if we would die. What irony that Bob Bickel, the brain behind that move at HP, is now our very own business dev head and has been instrumental in growing JBoss and HP is one of our partners today.


Open Source is soo hot right now. When I first started out, it was so not (JBoss started to mature as a product right when the tech bubble burst, and the disappointed expectations of the Linux Wave left a bad taste in everybody’s mouth). Everybody told me JBoss will never succeed: you guys are nobodies and you compete with every big player in this industry. Now the big players in the industry are falling over themselves to move into OSS or control it. The problem is that open source is about a lot more than marketing window dressing. Without street cred in OS development in your chosen space, without continued innovation on the product front and without highest-quality support from the core developers of that software, all you got is smoke and that ain’t gonna help you when the tiger leaps out of the underbrush.


I hope I have helped you grok some of the dynamics in our market. This is how I make sense of the recent moves in our industry. It is about a business model battle and it is reminiscent of the elephant and the tiger. Traditional IT superpowers, is open source your Vietnam? Time will tell.


Remember we love you,



PS: On Sunday as I was working on some of these blogs, my six-year old daughter came to me and said "Daddy, why are you working on a Sunday, is IBM still after you? (sic)." I laughed and said "Yes darling, IBM is still after me but now a whole industry is ganging up on us." She asked why. I thought about it a bit, how can you explain all of this to a 6 years old and answered "Because we might be small but we are mighty." She said "like mighty mouse!" She laughed at the image, herself and then, thirty minutes later, she came back with a rhyme and said "Here Daddy, include this in your blog and tell them it comes from me."


"Mighty mouse, mighty mouse
Is coming to your house
Be careful, he will trick you!"