Making Money In Open Source

Khalid writes "An interview with Sleepycat President and CEO, Michael Olson, it brings a lot of interesting information about their business model and licensing scheme. A lot of good ideas, when a lot of open source based companies are struggling for life. "

Comments

Making "money" vs "a living"

[j-beda]

OFten it seems to me as though when people talk about "making money" in Open Source (or in other contexts) they seem to be talking about "making shitloads of money", often with the idea of not doing much work.

Maybe we should refocus our attention on making "reasonable" or "sufficient" amounts of money. Statistically speaking, in most fields, nobody makes the bigs bucks.

In sports there are many many minor league players just scraping by for each major league bazzillionaire. In business, most places are small mom-and-pop (more than 50% of the US ecconomy?) compared to the relatively few McDonalds out there. Maybe Open Source is a bit different in that it is virtually impossible for there to be even this many (or any?) BIG winners - but that doesn't mean that everyone is thus forced to be a "looser".

It seems as though there are lots of opportunity for making a living with creation, support, etc. of Open Source solutions to various people's problems. And in fact there are a lot of people making decent livings providing those services.

You are right

[einhverfr]

Most OSS companies don't know how to run a business because the paradigm and rules of the game are too different. Look at Caldera, for example. There is a company which seems bent on selling Linux as if it were proprietary software, which looks like the safe choice to the novice, but really provides the worst of both worlds...

The real problem facing OSS companies is that the rules of the game are so different than they are for proprietary software companies. You cannot make your money on licensing because if you try to do that, your competition makes a cheaper clone. OSS commoditizes the software market.

I think that companies like Sleepycat are honestly trying to find out how to make money in a commoditized world, but I think they will have to transiton into another mode. Repeat after me: "You cannot make money selling open source software that you develop yourself." The market is simply too competitive.

What more companies need to do is offer tailored services, utilizing open source software to produce less expensive, more competitive packages of services. Of course, then you are competing with IBM, etc. but I suspect that this is EXACTLY why IBM is starting to move more open source.

It's a cultural thing...

[sterno]

A large number of americans have a lottery mentality that they are always looking for some way to make lots of money with little effort. All of the economic bubbles have been built on this tendancy. The dot com boom and the related boom in VC funding for open source was all about getting the big payment without doing the hard work. When the dust settled turned out a lot of those companies didn't have a clue and died.

I think there's a lot of money to be made on open source, but that's spread out over a lot of people. Support, custom development, integration, lots of useful stuff that isn't sexy and isn't going to make anybody fabulously rich but is valuable and will provide a more than adequate living for a lot of people.

I think if anything open source is a big bomb shell for the whole notion of making the big money. A lot of companies who got cozy making proprietary software, charging huge license fees and then selling exhorbitant support contracts on top of that are in for a big wake up call. When everybody can have access to the code you don't need to be addicted to one vendor. That breeds competition, and competition drives down prices. Those companies that can provide the best services for the best prices and can create the best brand will be the ones collecting the money in the future but because they necessarily must be efficient it won't be the really big money.

Software for its own sake?

[Phaid]

Maybe it's just me, but I don't really see how a company can make money at being purely an open source software vendor. The idea behind open source is that good software should be freely available to everyone. Selling software for the sake of software is sort of missing the point.

The company I work for develops embedded and realtime systems. Some are sold as products, others are custom developed for clients. We use Linux and other open-source software in our products. And we make plenty of money at it. But we don't make money from the software, we make money from the hardware and from our expertise at systems integration.

We support open source because we don't have to pay someone for basic things like operating systems. We contribute things like device drivers back into the open source community because they improve the OS, even if the drivers are for hardware which is too specialized for most people to care. Someone might have a need for this stuff and find it useful, and perhaps this someone might find a bug in it or add new features or expand it to be compatible with other, similar hardware.

And to me this is what open source is supposed to be all about. It's about people and enterprises with particular needs working together to solve common problems. As these problems are solved, the solutions get released back to the community and the software improves. Yes, this does make it rather difficult to simply be in the business of selling commodity software. In order to make money at it, you have to make the software do something useful.

And making software do something useful is what it's all about...

Fascinating, inspiring- I'm in the same boat

[Chris Johnson]

Thanks for the article, guys- what a nice thing for me to read right now. I'm fighting off an attack of what is probably tendinitis rather than, say, carpal tunnel, due to one too many 12 hour days working on _my_ software.

I do dithering/wordlength reduction software under the GPL (it's not C code, mind you, but it is still GPLed) and the last version did fairly well, competing with some extremely formidable proprietary dithers from places like the POW-R Consortium and Apogee. That stuff pretty well held its own for the most part with the best that the proprietary world had to offer.

Over the last week, I've drastically overhauled my noise-shaping code, and am trying to get final work done on a new release of my software- and this time, I've blown everything proprietary out of the water completely (w.r.t quality of output- not workflow convenience- I don't do hardware, or realtime output). I've got one dither that's consistently -160 db noise floor from 0-2K. Another one hits -170 db at 1.8 and 3K, right where the ear is most sensitive. Another steadily drops to below -170 db at the lowest frequencies. Another uses unusual methods to produce soundstage depth (a comb-filter-like noise floor- never heard of anyone doing that one before).

This is _all_ GPLed. You can't use any of it in proprietary software without violating the public license it's released under.

This is also _all_ mine. It's not even written in C, much less based on GPLed code from others. (That's one reason why my file reading routines suck ;) ). It uses none of the 'many hands make light work' concept of open source- I do all the work and have not had any offers of help from anyone at all, except some mastering engineers who've given their thoughts on early versions of its audio performance, and they don't write code.

With that in mind, I have to say I'm delighted to see SleepyCat's take on all this: it confirms what I had suspected, and gives me hope for the future. You can make money on open source by indulging in people's desires to NOT play along and open their own source. You can charge them to NOT share (if they share, they get it for free. How much is your paranoia worth to you? ;) )

The only requirements are that YOU have to do the work- which stands to reason- that you have to not only do all the work but also outperform everybody else- and that there have to be enough others out there who want what you have to give, but won't themselves share. Basically, you're charging people for their own greed. If they were willing to give their work to the world as you do, they'd get your help for nothing. But if they want what you have, and won't do as you do- they must either do without, or come to terms with you. (or rip you off outright, but that's another story).

It's inspiring to see how these folks do it, and definitely something to emulate- makes me glad I've been using pure GPL all along, rather than something like LGPL. They're so right- you can't exert this kind of pressure UNLESS your 'free' licensing is hardcore 'libre' with no concessions to business. It's gotta scare the suits and the lawyers enough so they come to you and say 'maybe we can work something out' (*kaching!* good to do business with ya!).

Wonderful to see this. And again it's so simple that I marvel that nobody else has been suggesting it- I thought I had sort of invented this concept for myself out of necessity and it's reassuring to see that people have actually tried it and it works. You have to have a product so good that people _do_ want a piece of it- a libre license that scares the suits- and a willingness to release private versions under non-free parallel licenses to companies that want what you have, but won't share code themselves.

Maybe this _does_ lead to doing yourself out of a job, in the future when everybody is nice and shares ;) or maybe the 'bar' is very high, in that you really have to perform to get in such a position. If that's the case, then (a) explains why I haven't made money this way yet, and (b) if 20 db better noise floor than the top proprietary dither isn't enough, I'll keep working until it's 30 db ;) it looks like some entire concepts like indeterminate-order noise shaping are mine alone. I don't think you can even get error distributions anything like what I'm getting without it- so for a change, rather than indispensable technology being owned by a patent holder, indispensable technology is 'owned' by the sphere of GPLed free software. Any arguments that code or algorithms are property and not speech will only enhance the value of this 'ownership'. Or to put it another way: so software is not speech? So you can 'own' an algorithm, huh? Well, _this_ one you can't have unless you go libre with your own code, or pay! *kaching*

Hah!

I gotta get back to work- the main dithers have been hammered out, but I need to adapt some of the others, like Logic and Ambient, to the new error-feedback routines. They probably won't outperform the others in numbers, but they're geared to different needs: Logic turned out to be good at revealing depth cues accurately, and Ambient was particularly warm with very authoritative bass, and I gotta see if I can bring out those qualities more.

If there's anybody else out there capable of doing this type of thing in other fields, please, please, go for it with every bit of effort you can come up with! Maybe 'open source' as a way of not having to do as much of the work has an immediate appeal to people- but the only way we're gonna REALLY get Software Libre out there and impossible to avoid is if you do the work yourself, do it BETTER than anyone else can (pick your field carefully, and narrowly!) and then put your stuff out with a hardcore libre license and a willingness to charge for dual-licensing! It's gotta be 'This is mine- you either share, or you pay. My way or the highway'. That's the only way to win...