Buddy Dirk Riehle from SAP research has published an interesting paper on The Economic Motivation of Open Source Software: Stakeholder Perspectives.
I agree with his approach that only an analysis of each stakeholder's specific perspective can identify the different incentives that drive the economics of open source. The basic underlying assumption is that of an homo economicus, which I share, too.
The interesting part is that it is overall a win-win situation (except for the monopolists in the pre-open source era with excessive profits, particularly Microsoft, Oracle, and his own firm, SAP).
Why is that? The reason is in the power of collaboration and the fact, that software is an immaterial asset, which can increase its overall value through sharing. Unlike in the material world, where either you or me can have an item (e.g., a new Macbook), immaterial goods can be copied and shared, giving value to many people. Shared music is the example easiest to understand (if you share it with me, we can both enjoy it equally), but also a shared idea or software code can provide value to different people in a different context, which makes sharing immaterial goods a larger-than-zero-sum transaction.
Another way to put it: Sharing immaterial goods like software makes the pie bigger.
Dirk argues that the distribution will shift gains to one stakeholder at the expense of another.
I disagree. My thesis is that the pie will get bigger and that everybody who embraces open source will gain, thanks to efficient markets and some forces that Dirk hasn't analyzed sufficiently, specifically:
- System integrator
I think system integrators find value in software vendors giving them standardized software, procedures, a roadmap, training, leads and much more. If SIs work with community open source much of that value goes away, the SIs have more cost. So it's not just pure profit for them here.
As an example: IBM has hundreds of people in the Linux group, writing the software that they will support later. Yes, Linux is license-free for them, but it's also more costly to support than a very directed piece of software like, say, Windows NT, which comes ready-to-eat.
The other element is that those gains will not remain as excessive profits. Economic theory predicts, that this very dynamic market will pressure the total price point down to "normal profits", which should be much the same as before.
- Software vendor
I think Dirk misses the point that, with open source and the Internet, the competitive barrier doesn't come from IP protection (which articifically creates a "material world" situation of non-shareable goods), but from building a strong user community.
Dirk and I will know from the economics classes in our alma mater, that software vendors live from network effects, which stem from increasing marginal returns through additional users. The big change in thinking is that we used to measure those network effects through market share in license sales, because they correlated with the installed base.
That's not true any more. With open source we start to count USERS as the determining driver of network effects.
Toplink, Oracle's ORM, for instance, had more sales$ than Hibernate, but much less users. It lost (they dumped it into the open source space)!
MySQL is commercially a <1% dwarf compared to Oracle, IBM and MSFT's RDBMS business. But with its installed base and users, it has already reached 1/3 "market" share, harnessing huge network effects (brand equity, ecosystem, horizontal applicability, etc.).
With that competitive barrier, MySQL can successfully fend off community open source projects (like PostegreSQL, HSQLDB etc.) as well as effectively attack the incumbents. - Employee
I absolutely share Dirks' analysis about the career building aspect of open source work for software engineers and I have spoken and blogged repeatedly about it.
I think he misses a major point, though: With the open source production model, properly implemented like at MySQL or db4objects, developers work from their home and communicate over the Internet, no matter whether they are paid or not. As a result, companies can recruit from all over the world and leverage a global talent pool rather than a local one.
I have talked about this what I call open source globalization in several posts and articles.
As a result there is no linear payoff table for employees like Dirk suggests.
It rather looks like this:
If I am in a low living-cost country, it's just gonna be great.
If I am in a high living-cost country, it will depend on how much I will be able to differentiate myself from a "commodity coder". Becoming a committer, as he suggests, is certainly one way. But it will be more competitive day by day. I think a better differentiator for an engineer is to focus on being close to customer and markets, i.e., a location-based differentiator. This could be some evangelizing work or customer projects/interaction, for instance. Of course, doing politics at US based open source project meetings to become influenial, is another way of doing it, though it defies the open source meritocracy aspect.
It is quite revealing that Dirk talks about employees. I rather think of paid contributors as mini-entrepreneurs and, at db4objects, they are technically contractors.
So I believe that open source makes the pie bigger and that everybody will win, who understands this "post-material" world. I actually disagree that the payoff will be radically differently distributed, if each stakeholders adapts accordingly and keeps providing the value they do. Efficient markets will take care of this.
The loosers will be the laggards, those that don't understand and don't embrace the model.
Those are my 2 cents.
Have a great weekend!