"FOSS Business Model Broken" — Former OSDL CEO

liraz writes "Stuart Cohen, former CEO of Open Source Development Labs, has written an op-ed on BusinessWeek claiming that the traditional open source business model, which relies solely on support and service revenue streams, is failing to meet the expectations of investors. He discusses the 'great paradox' of the FOSS business model, saying: 'For anyone who hasn't been paying attention to the software industry lately, I have some bad news. The open source business model is broken. Open source code is generally great code, not requiring much support. So open source companies that rely on support and service alone are not long for this world.' Cohen goes on to outline the beginnings of a business model that can work for FOSS going forward."

Well, duh

Do IBM sell software? No, they sell you a solution. They'll sell you Linux, AIX, Solaris (IBM is Sun's second-biggest seller after Sun themselves) or Windows.

Don't sell "software", sell "a solution to the customer's problem." This sounds cliched, but it's amazing how many people and companies work around actually doing so.

Re:Well, duh

Don't sell "software", sell "a solution to the customer's problem."

I think that's key. The point in TFA was "open-source code is typically of such high quality that you can't make money selling support". While I htink that's a bit self-serving, as there's *planty* of crap open-source code out there, it's almost universially true that open-source code that lots of people use converges on "so high quality you can't sell support" over time. (This as opposed to commercial software that sometimes gets it exactly right, but then goes on to break everything in the next release because you *have* to have a next release.)

IBM does very well selling consulting services. "Open source" is a nice way of saying "we're going to take the code you pay us to write and use it to solve the next guys problem too". And of course that works out well for everyone, since this customer benefits from all the previous companies. Cunsulting firms do that *anyway*, of course, but calling it "open source" gets it all above board *and* lets unrelated people benefit.

Re:Well, duh

Do IBM sell software? No, they sell you a solution.

Actually, they do. $19B dollars' worth. That's a lot of software, or about 20% of their total revenue (for 2007, at least). Services (or solutions) revenue is tracked separately.

Granted, the software IBM sells is a solution to a problem. But to say that IBM doesn't sell software is like saying Gap doesn't sell clothes; they sell a solution to the nakedness problem.

Re:Well, duh

What nakedness problem?

Not a Factory

Don't sell "software", sell "a solution to the customer's problem." This sounds cliched, but it's amazing how many people and companies work around actually doing so.

Yeah, this guy seems to think the the 'support' model equates with the kind of technical support he might get on the phone at an 800-number.

In my business, anyway, the open source support I sell is really business support. Companies want to know how improve their business with software, and I can help them figure that out, and open source is most often the best answer. I usually save them a bunch of money, deliver a robust solution, and pay some bills by doing so.

Granted, that's not what most 'investors' are looking to do - they want to mass-produce support scripts for that 800 number and charge $40/call. But in my case, what people are really buying is my ~20 years of IT experience and knowledge and its application to cutting-edge technology, which can't get mass produced by the end of next quarter.

The broken business model is applying factory thinking to knowledge work.

Re:Well, duh

Then the article's saying "too bad for them." Proprietary boxware is on the way out. Proprietary vertical market stuff gets toward "solution" selling. (Certainly at the prices they charge. Honestly, the more it costs, the worse it appears to be in quality ...)

Re:Well, duh

"Honestly, the more it costs, the worse it appears to be in quality ...)"
Not really. Your looking at the per seat price.
Let's take two examples. One is a program for managing say a U-Storage location vs Office.
The U-Storage vendor might charge $2000 for their program while Office costs say $295.
The U-Storage software vendor might sell 200 a year. Microsoft will sell what? Make it a million so the Math is easy.
So the total cost of Office to the Planet is $295,000,000 while "expensive" vertical cost only $400,000. That might seem like a lot but then you have to think about marketing, paying the Programming staff, and support costs which will be much higher per customer for the vertical.
That expensive is software is actually very cheap when you look at the that way.
That is the advantage of boxware. Buy charging for every copy you can spread out the cost of development.

Re:Well, duh

That is interesting. I have not seen that kind of error in forever. I did put that kind of error into one of our in house programs intentionally.
We had a clerical person that kept putting their machine into 640x480 so she could have big letters on the screen. Well my program just couldn't work well in that small of a screen. No matter how many times we told her to just pick a bigger font she wouldn't do it.
So I got tired of her complaining about my program so I just put in a check and the program put up an error message if the screen wasn't at least 800x600 and refused to run.
She never seemed to get it that I made the program do it but she obeyed and set her screen to a better resolution.

Re:Well, duh

Do know that it is generally estimated that less than 20% of programmers work for software companies?

Lots of companies have in house programmers. I have even seen cities with "hiring freezes" post ads for programmers and sys admins.

Every estimate I have seen for the failure rate for computer projects is 80% or higher.

More or less everything with computers has been a disaster.

Bespoke solutions, off the shelf solutions, a combination of the two, you name it, your work will be harder and the money spent will never be recovered. (at least that is the way to bet.)

Of course we see the relatively rare successes and assume that the massive carnage was just the unenlightened fools that didn't know what they were doing.

Here's a great paradox for ya..

how supposed "experts" can be so dumb.

support != hand holding.

All software has bugs. If your customer finds a bug in the software they can report it upstream and wait around for the bug to get fixed or they can report it to you and pay you to fix it now. That's support. Same goes for features. Maybe they want to use the software for something that upstream thinks is worthless. They could beg upstream to add the feature. Or they could hire developers to add the feature. Or they could outsource that to you. That's support.

Re:Here's a great paradox for ya..

Wow. The commercial software mindset really is taxing isn't it? You don't push custom features upstream.. upstream won't even accept them unless they are something everyone would want..

Re:Here's a great paradox for ya..

I think the solution is to push for a plugin system in the upstream. I makes downstream customization easier for everybody.

Re:Here's a great paradox for ya..

As long as features are being added, bugs will occur. In a sense they are an infinite resource, since I can't think of much in the way of commonly-used software with no feature development. Hell, Windows XP has been "feature frozen" for years now, and yet I still get updates. Or - to use a non-MS example - Python 2.3 went final in the summer of 2003, and yet there was an update this year - almost 5 years after it was "feature frozen". (For reference, Python is up to 2.6 - and their 3rd RC for 3.0)

Re:Here's a great paradox for ya..

As long as they do their job then there is no need to update them.

No question, but the world changes around these machines. In an industrial setting I've seen a couple examples of this...

One example is a computer that controlled a visual inspection machine that ran Win 3.11. Worked great until Y2K! In addition, the newer hardware doesn't always run that version of Windows and it was becoming a problem. Additionally, the company was changing to a new file server that Win 3.11 didn't like. Eventually, the whole big thing had to be replaced because the stupid little Win 3.11 machine didn't cut it anymore. Had Win 3.11 been open source and a consultant was able to change it, the machine would probably still be operating.

Another example is at a plant where all of the flow control was done through old DOS programs hooked into serial-based equipment. These days, you simply can't buy the serial-based stuff any more... everything has ethernet. Additionally, it has become very hard to find modern PCs that will talk to the old hardware at all. As a result, all of that old DOS stuff is obsolete and being replaced. If the old DOS software was open source, it could just have been hacked to add ethernet support.

Anyway, my point was just that the tech world changes pretty fast, and a business may THINK it wants stable, unchanging software... but I bet that isn't the case. After all, when's the last time you saw a text-based ATM?

Why does nobody understand why this doesn't work?

Open Source development is an absolutely amazing and powerful tool... ...for everyone in the world whose livelihood comes from something *other* than selling software.

Bankers need to run their banks more efficiently so they get together to cooperatively develop some banking application software that makes them all work more effectively and efficiently. This is the magic of co-op software development. There are other people who have the same problems you do, and if you get together you can produce really useful software for vanishingly small cost, and the result can be replicated without limit or expense.

Bankers don't, on the other hand, create free, zero-income banks.

Commerceial software companies making free software is, and always has been, a really dumb idea.

If you find yourself in this position, my suggestion is to move up the food chain towards applications of the software you've developed. Eventually you'll find a level where people have problems they're willing to pay to have solved because they're not common enough to make an open source / co-op solution viable.

If your business plan reads:

1) Invent really cool new product.
2) Give it away for free.
3) Enable the community to do all their own support and enhancements.
4) ????
5) Profit!

let me save you some time and point out that there is nothing you can put in step 4 that leads to step 5.

Open source development is not a segment of the software indusrty, it's a segment of the every-other-industry.

G.

The FOSS Business model

The model of investor expecting to make a quick buck off FOSS is broken. Not FOSS.

Re:The FOSS Business model

Finally! I'm so sad that I had to go halfway down the page to get at this one...

Working with FOSS isn't going to lead you into a business hole. Businesses still need someone who can monitor, upkeep, fix, and add to the machines that power their business. If you happen to write FOSS to do those things, you can make money by still doing work for said customer. If you merely use tools that are out there done by other coders, then you are still providing a service for the business/customer.

The break in this article (and many like it) happens when said supporting company goes "public." Suddenly, profit margins must be maintained, P/E ratio enters the equation... Quarterly performance is the measurement of your business, not "am I doing a good job?" FOSS has its place in business, and in business models. You can make it your job to contribute to the community, or to utilize it, to help generate cash for the business that runs on it... Just don't expect your stock to split anytime soon.

Something that I've learned: The business always has needs. They might not know what they are, and they will be different in 6-12 months. The needs will always exist, they will always require someone to implement them, and then maintain them... FOSS often provides a solution to those needs... Even if the code could be perfect, and you could, theoretically, never have to maintain that FOSS solution, you'll be needed to implement someone else that the business now needs.

Great code NOT EQUAL TO ease of use

Open source code is generally great code, not requiring much support.

The code itself might be great, but generally, the front-end (which I'm distinguishing as separate from the back-end nuts and bolts "code") is a mess. Installation and use difficulties are generally greater in randompackageX off of SourceForge than, say, MS Word or FoxIt. There are some OSS programs that are near hitchless, like Pidgin or Firefox (had noticeable problems with crashing on exit in Vista, though), but if you go beyond the star players, you'll quickly find this argument doesn't hold up to empirical scrutiny.

Traditional?

...the traditional open source business model...

It's been around long enough to be "traditional"?

"Open Source" is not a business model

The open source business model is broken.

It isn't a broken business model. It isn't a business model.

Saying the open source business model is broken is like saying open source doesn't work as a cheese sauce. It also isn't a very effective screw driver. On the other hand, I have yet to hear a business model you can dance to.

--MarkusQ

Solution in a mixed model?

As far as FOSS being something that has serious business problems in regarding to sustaining the developers who work on it, this is indeed a serious problem. It generally can be very hard to raise revenue with FOSS, projects can ask for donations and sell packaged versions, but you often end up with just a trickle with these sorts of things. Programmers should obviously be able to work full time developing software. With FOSS directly competing with commercial software an eroding those markets, could it be that programmers will end up waiting tables during the day just to support the time they spend writing code? fOSS does indeed wipe out commercial software markets and it can actuall

I am supportive of the freedom aspect of FOSS. For far too long commercial software has shut down innovation and stifled the development of improvements through cooperative development with its closed model. FOSS is on the other extreme, its an open model but it leaves programmers in a situation where they cant afford to live. Perhaps a solution for some projects lies in the middle, with a commercial source tiered licence system, where the source code is provided with all licences, the developers are receptive to improvements from customers, and the cost of software is set according to the ability of the customer to pay, a hobbyist who is using the software for fun would pay far less than someone using it in a high revenue business. This assures that the software does have a high degree of openness and accessibility to all, but also assures revenue can be raised to develop the software.

Don't give it away for free

One thing i think we will see FOSS project's movng away from is giving away the software. if you GPL something, it doesn't mean you have to give it away, it just means who ever you sell it to gets the source code along with the program.I could for example write some software, sell it to others and then give them access to the source where only paid customers could make commits and see the source. source is only required if you distribute something....

Re:Don't give it away for free

No, you can charge as much as you want for distribution, and you only have to give source to those to whom you, personally, have distributed binaries:

http://www.gnu.org/philosophy/selling.html [gnu.org]

However, as you noted, all it takes is for one customer to put the source up for download, and you're screwed.

A bio is useful here

Stuart Cohen today is chief executive officer of the Open Source Development Labs. With more than 22 years of international sales and marketing experience, he is a seasoned technology industry executive and has served in a variety of executive roles. Most recently, Cohen was vice president and corporate officer at RadiSys Corporation where his responsibilities included strategic partnership development with other industry leaders including IBM, HP and Dell. Prior to RadiSys, Stuart was vice president of worldwide marketing and a corporate officer at InFocus Corporation. Stuart spent 17 years with IBM, where he held senior positions in the US sales & marketing division, and the IBM Personal Computer Company and Networking Division, with international business development responsibilities in Europe, Southeast Asia and China. Stuart holds a B.S. in Quantitative Business Analysis from Arizona State University.

LinuxWorld [linuxworldexpo.com]

So... a corporate marketdroid that never invented anything, never built a business, who coattailed himself into executive positions with minor players based on prior employment relationships with major players who has a B.S. in Quantitative Business Analysis. Who, coincidentally is trying to bridge free software and services in a for-pay model that's starving for attention [csinitiative.com]?

I'm gonna go with... um... so they couldn't get an Enderle quote? Was Maureen O'Gara busy that day? How did this guy talk his way into OSDL? It's interesting that their Wikipedia page [wikipedia.org] mentions him not at all.

That's great

So long as you are content to have a number of kinds of programs never be open source. If your solution is "Bundle with hardware if you want OSS and need money," ok then don't be surprised when people who's market doesn't deal with hardware chose money over OSS. I'm talking about things like games, or, say, video editing software. Things where you neither want or need additional hardware. Things where the idea is to use the hardware of a general purpose computer to do what you want.

This accounts for most software out there. While there's certainly things like, say, a firewall app/OS or something that it is perfectly valid to bundle with hardware, there's plenty of things that are just programs to run on a normal computer, no other hardware needed or wanted.

For programs like these, the response from many OSS advocates has been "Sell support!" However that doesn't work in a lot of cases. If you program is well written and easy to use, people won't need support by and large. Some of my favourite software packages, OSS and commercial, are ones where I don't need support of any kind. They do their jobs and are easy enough to use I need to additional help beyond what's included.

So what then? What do you do if your software is both a good product, and not one that uses hardware? Currently, the options seem to be "Open source it and give it away for free," or "Close source it and make money." In some cases, people can afford to do the former but not all. The "Just give it away for free," sounds like a nice idea when you are a broke student who would be receiving said free software. It sounds like less of a good idea when you are a programmer with a family to feed who would be getting no paycheck if you do.

So you run in to a large category of programs where you don't have a viable model. Support isn't a viable model since people don't need it. Bundling isn't a viable model since that isn't what your software is for.