Simon Phipps, Sun’s Chief Open Source Officer, maintains that the software industry is gradually moving from a proprietary business model, based on the End User License, to another based on Open Source. Now, you might be thinking “With a job title like that, he’s gotta peddle that idea”, but bear with me. First of all Simon is pretty much an FU person, who says what he thinks. He did so when employed by IBM as a Java evangelist and has done so since joining Sun, where he has migrated from Java to Open Source.

I spent most of yesterday afternoon talking with him at a pleasant Austin Starbucks. We talked for so long about so much that I wish I’d been recording it, but the most interesting exchanges were about Open Source.

Sun and Open Source

Sun, believe it or not, intends to gradually release all its software products into Open Source. It’s already got a number of important products in Open Source including the Solaris OS, Open Office and the recently acquired My SQL. According to Simon, Sun will move more products in that direction as it changes its software business model.

Simon explains this model very well. Traditionally, enterprise software has been sold by software sales reps on the basis of a one-time fee and recurring maintenance revenue. PC software has been sold on the  basis of a one-time fee and upgrade fees, which is a similar model. In either case you get the situation where a prospect pays a fee to be transformed into a user. Consequently you get a business model based upon the transformation of prospects into users. That means a sales force, advertising, marketing campaigns and so on. The cost of adoption is high on the user side as well because the high-ticket price necessitates procurement procedures that can be long and involved.

The whole process is expensive for everyone concerned and it also determines how the software evolves. The source code is a deeply kept secret. Actual changes made are at the vendors discretion and it is in the vendors’ interest to keep the software evolving whether it needs to or not. So ultimately every penknife gets transformed into a Swiss Army knife with multitudes of features, including that cant-do-without-it-metal-prong-for-removing-stones-from-horses-shoes.

The alternative model that has evolved through Open Source is distinctly different. It is based upon turning users into customers (and collaborators).

Let’s deal with the downside to this approach first. The downside is simply that the revenue stream comes later and it will probably be lower. The normal evolution from user to customer occurs when the Open Source software is being used in important applications and the user company concludes that it should buy support. From the customers side this is a wonderful deal because you never pay for software until it has proved its usefulness. There is another benefit to the customer. None of the smoke-and-mirrors software sales antics go on. You can speak freely to other software users and get a reliable view of their use and appreciation of the software.

The vendor is stuck with the considerable issue of developing revenue streams, but the revenues streams need not be as large as with proprietary software because there is no expensive sales force and marketing activity. No law proclaims that you cannot charge a one-time fee for an Open Source product, but doing so creates a blockage to take-up, because it means there will be a procurement process (even if it is a trivial one because the price is low). Free adoption is important. The main revenue stream is thus likely to be support and consultancy - although in some instances advertising can be used to generate revenue.

The difference in how software evolves with Open Source is dramatic. This has become especially so in recent times when Open Source products, like FireFox and WordPress for example, have published a plug-in interface that allows other developers to develop plug-in additions to the basic software. What can happen (and has happened in both these examples) is that a whole software ecosystem develops around the base product. When you understand this, it no longer seems remarkable that FireFox has such a lead over Microsoft’s Internet Explorer

Sun, incidentally, is now ensuring that Open Office follows such an evolutionary model. Instead of feature overload, feature selection by plug-in. It’s a particularly refreshing idea for word processing.

An Alternative Perspective

As a rough rule of thumb the cost of a manufactured electronics product will fall by about 30% if its sales double. Naturally much depends on raw materials and the manufacturing process, but the underlying point is simple. Large volume delivers economies of scale allowing you to invest in improving the manufacturing process itself and strike better deals with suppliers. Also the front-end set-up costs (design costs, prototyping etc.) are only paid once and don’t recur.

The same should be true of software. Particularly once a product, like Windows for example, is out there, its cost can reduce quite dramatically with volume. The cost of distribution with such a product is minuscule. And yet we have seen Windows at the same price or even higher price as time has passed - and this has happened while computers were increasing in speed ten-fold every six years. Windows is only marginally better after each such six year cycle while the computer it runs on has been completely transformed. This has happened only because of the power of the Microsoft monopoly.

The truth is that all of the activities involved in building software since the advent of Windows have fallen dramatically.

  1. Developers cost less (because the world has opened up and developers, in many Open Source projects, will work for pizza and options).
  2. Development tools cost less (most of them cost nothing).
  3. Powerful computers cost less (all the better to test them with)
  4. Communications cost less.

This, of course, is why Open Source software is viable. And it is also why the proprietary software model is challenged.

Whether the timing is right for what Sun is doing is open to debate, but I’m in agreement with Simon as regards the sense of it. The software industry will move to this model more and more as time passes. It has to.

  Subscribe to HaveMacWillBlog in a reader