June 01, 2005

Software Licensing and the Know-how to Issue

Software charging for big ticket sellers is getting more complex again, as dual cores from AMD and Intel start to invade the small end. Oracle, which made billions charging on the muscle power of CPUs, will have to do something, and we've by now all seen IBM's adverts on TV suggesting "on demand" with its concommitant charging suggestion: You demand, we charge.

I've done a lot of thinking over the years about how to licence big ticket items like issuance software. In practice it is very difficult, as the only revenue model that makes sense for the supplier is for large up front licence fees to recover large up front capital and sunk costs. But for the demander (issuer and user of the software) the only model that makes sense is to pay later, when the revenues start flowing...

Issuance software has all the hallmarks of an inefficient market and I don't think there has been successful case of issuance licencing yet, as those two "sensible" options do not leave any room for agreement. This may be rational but it's very frustrating. Time and again, we see the situation of people wanting to get into the issuance market who think they can produce the software themselves for a cheaper price. And they always end up spending more and getting a lesser quality product.

In practice what we (Systemics) have been doing is this: running the software ourselves as "operator", and charging operating costs, with some future licencing or transaction flow revenues. Yet, the deal for future revenues is always based on a promise and a prayer, which is already asymmetrical given that most startups do no more than start up. (And it isn't just me bemoaning here - if you look back through history there are literally hundreds of companies that tried to build value issuance and sell it.)

Which leads to the freeware model. In the freeware world, big ticket items are given away and money is made on the consulting. This has worked relatively well in some areas, but doesn't work so well in issuance. I'm unclear of the full reason why open source software doesn't work in issuance, but I think it is mostly the complexity, the sort of complexity I wrote about in FC7. It's not that the software can't capture that complexity but that the financial cryptography business often finds itself so squeezed for management complexity that partnering with a strong software supplier are beyond capabilities.

What will potentially help is p2p issuance. That is, "everyone an issuer." We've always known this model existed even as far back as 1995, but never really considered it seriously because too many questions arose. Little things like how we teach grandma to sign a digital contract. We've now done enough experiments in-house to confirm that the corporate internal issue and the individual issue are workable, sustainable economic models but we have to get other companies and individuals to do that and for the most part they still don't do anything they don't understand.

I'm guessing the way forward here is to turn client software into issuance software. This brings up a whole host of issues in financial cryptographic architecture. For a start it can never seriously scale simply because people do silly things like turn off their laptops at night.

But, more and more, the barriers to issuance and financial cryptography in general I believe are spreading the knowledge, not the tools and tech. Every year our tools and tech get better; but every year our real barriers seem the same - how to get users and customers to make their first tentative issue of a currency of value. Oh, and how to make money so as to keep us all alive, which was the starting point on this long rant of liberal licence.

A couple of footnotes: In a similar thread over at PGP Inc, Will Price reveals how they've managed to get out of the legacy freeware version trap:

"When the 30 Day Trial version of PGP Desktop Home expires, it reverts to a set of functionality comparable to what used to be known as Freeware, and said functionality remains available indefinitely -- under the same license conditions as Freeware used to be under."

Nice one. That works for client software, not for server software.

Here's a further article on how the big companies are also working out how big ticket software isn't the way to go:

Posted by iang at June 1, 2005 09:48 AM | TrackBack
Post a comment

Remember personal info?

Hit preview to see your comment as it would be displayed.