I'm plowing through the paper that Nick Carr introduced, and finding it fascinating. The paper, as mentioned before, is concerned with how we measure the value of free users of a product.
Consider the case of Monster.com, an employment market place where job-seekers post their resumes and firms sign up to find potential employees. Monster provides this service free to job-seekers and obtains revenue by charging fees to the employers. A natural question arising from this business model is how much Monster should spend to acquire a job-seeker. Traditional models of CLV can not answer this question since job-seekers do not provide any direct revenue. In fact, if one includes the cost of maintaining resumes, the standard CLV [Customer Lifetime Value, or net present value of all future profits from a customer] for a job-seeker is negative. However, without job-seekers employers will not sign up, and without these firms Monster will have no revenues or profits. In other words, the value of job-seekers is through their indirect network effect on job listers. This indirect network effect is not limited to employment services only (e.g., Monster, Hotjobs, Craiglist) but also extends to any exchange with multiple buyers and sellers (e.g. eBay, real estate).See the issue? A company starts with the reasonable, but too-basic, revenue model of "We make things and people buy them from us." Such a model, however, fails to account for the world of Web 2.0, open source, etc. that we find ourselves in.
In this new world (just as in the old, though perhaps more pointedly now), there is tremendous value in "free riders." But how do we determine how to treat free users of our products?
One statement in the paper that jumped out at me probably qualifies as a platitude, but it has interesting implications for an open source company. It hold some promise for answering my question above:
As a firm acquires more buyers it becomes more attractive for sellers to join the firm as well. The reverse is also true — the more sellers a firm has, the more buyers it is likely to attract.Why can eBay charge sellers so much? Because it has the most buyers. Same with Google (except, in Google’s case, it brings passive buyers, or eyeballs/pageviews/click-throughs). Not raised by the paper, but which also applies: why will buyers often pay a premium, or at all, for a product? Because it is in use by so many other buyers. (Safety in numbers.)
The question for a vendor, I suppose, is whether our free users are buyers or sellers. Or, more pertinently, which do we want to make them, because I believe an open source vendor can do either.
If we treat them as buyers, then our goal should be to maximize conversion rates, closing down as much free-riding as possible. If we treat them as sellers (or, in other words, developers), then our goal should be to maximize the attractiveness of the platform for them, such that they derive some value (whether reputation, money, whatever) from “selling” their services into our project. (This is one of the core reasons for Alfresco's move to the community-friendly GPL.) Or maybe we allow them to actually make money as partners (the way SugarCRM does on the SugarExchange), and derive direct or indirect value/money from that activity?
The point is that where we go depends on how we view our community: buyers or sellers. What seems clear is that we can’t treat them with indifference. We have to actively mold it into one or the other, or (more likely) some combination of the two.
I don't believe in trying to upsell a community, as Fabrizio notes). I doubt Google's community of free users would embrace a model where you could get basic search for free, and enhanced search for a fee. Google works because everything is free until you actually think of buying something (i.e., click-through). You are charged for acquisitiveness (which is the opportune moment to draw dollars out of the experience).
In open source, there is tremendous value in having a large community of users, paid or unpaid. Some of these users contribute cash. Some contribute code. All contribute safety.
There is also value to Microsoft. Think about its current patent offensive. I would argue that Microsoft would be better off (i.e., make a lot more money directly licensing its patents) by setting up low, publicly-priced "rents" for its patents. Instead of closed-door bartering over what they cover and how much they're worth, let people "subscribe" to patent coverage. Maybe, to encourage, adoption, let end-users have them free of charge (meaning, free of charge beyond the cost of the product itself; in other words, eliminate litigation FUD), but charge vendors. (We will not sue customers, but we will sue vendors that violate our patents.) I'm not sure - just "thinking out loud." I believe Microsoft would derive more value by encouraging (some) free use of its property than trying to extract absolute value from its patent portfolio.
As for open source, Compiere's model already subsidizes end-users of its ERP software (it's free) and charges partners, instead. Maybe this, or a variant of this, is the right model?
Or maybe Red Hat's model, which gives the software away for free but which charges for services (support, Red Hat Network, etc.)? Red Hat treats end-users as both buyers and sellers (developers), but uses Fedora and RHEL to effectively distinguish between the two.
One thing seems clear to me: treating all free riders as criminals is almost certainly the wrong model for future growth in the software market. As you can see, I don't know what the right answer is. But any successful company in the world of software and the Internet must depend upon a healthy dose of free users. Otherwise, it's unlikely to get paid. Think about that. :-)
Posted by Matt Asay on March 8, 2007 01:04 PM
