Tony Mobily's recent FSM post A future without Microsoft and the resulting comments have caused me to consider the way we use numbers to argue for free software in the marketplace. I'm not convinced that it's the best strategy because those waters are particularly muddy when it comes to comparing free and proprietary software.
I suppose it's not always a good idea to criticise your editor but then again I think I know my editor and he can take it. In truth Tony didn't go too deep into the numbers game so I think he survived unscathed. Anyway as Lando Calrissian said, here goes nothing.
Play our game not their game
Regular readers of this column may know I am an advocate of making an argument on the merits of your product and not the faults of the opposition. In the same way I think playing the numbers game with Microsoft is a foolish thing to do: that is their game and they are good at it. Despite what they say they are Microsoft are--at heart--bean counters: their bottom line is--well--the bottom line. What they care about is how many units they have sold. They are experts at manipulating statistics to show what they want you to see. For them it's all about the numbers and they are not alone. Proprietary software companies across the globe seem to follow daintily in the footsteps of the Redmond giant (sometimes if they get too close they get squashed but that's a different story). Their mantra is "business is war" and in war the one with the most people standing at the end, wins. In software the one with the highest numbers wins. It's rare to see free software projects actually aiming to destroy their competitors--it happens but not as much as in the proprietary world.
The free software community is not that kind of animal and nor should it be. Whilst individual businesses may employ free software, the software itself is not really about winning, it's about doing the best job it can whilst maintaining the freedom of its users. Numbers of users and installs are only relevant if those numbers mean something. What they mean to proprietary companies and what they mean to a free software project are poles apart. In the first instance it's about revenue, in the second it's about how many people you have supporting each other. Both are independently factors in the success if the product but the nature of that success differs vastly between the two models. Software sales figures are about revenue from what you did. The game has changed, it's now about revenue from what you do, not what you did.
A new game
Free software didn't change the game so much as revert it to its origins -- before it got corrupted. It then built upon the original game and -- if you like -- forked a new branch away from the proprietary control-freak one. The game is one where numbers no longer count for much. Saying there are 20 gazillion users has always said little about your product. A proprietary product sold to a customer does not mean it is still being used or any good. Equally and free software product downloaded does not mean it is still being used or any good. A proprietary sale does mean a single user/instance but a free software download does not. You just compare those two sets of numbers and as a lot of users find out: popularity in either camp does not mean suitability to your particular need. If we are ever going to break the "nobody got fired for installing X product" mentality then we need to stop harping on about numbers because they mean so little in reality.
End-users, IT managers, CTOs, CIOs and any other person with information or technology in their job title need to wake up and smell the coffee. They need to see that they've been sold duff advice. Buying the "biggest seller" does not necessarily mean increased support and certainly does not automatically mean greater suitability. It means you are investing in somebody else's business, not yours. Once you leave the sales and marketing game the picture becomes a little clearer. You can focus on what the software does, how it integrates, what kind of support you get, how does it perform. And when I say focus on those things, I mean you get to find out for real as opposed to a sales person promising what their developers can't deliver on.
Free software is rarely sold and if so it's even rarer to find it's the licence that is being purchased. Proprietary software is almost always sold, free and bundled version s appear to contradict this but -- as most of us know -- you pay for it in other ways. Comparing two products which propagate in entirely different models is next to useless. You always have to try and put one into the other's terms and it usually doesn't fit. I've said this before in a different context but it's like comparing apples and air, and it's as big a waste of time if you ask me.
Comparisons
The problem with my argument above is that the other lot keep comparing in the terms of their numbers game and those are terms the end-users is familiar with. So how do we combat that? For a start we can argue on our terms, within the new game. We can also demonstrate the effectiveness, not the popularity of our products. Finally we can do all we can to highlight the fact that the game has changed and that the numbers game is going to be one that smart people no longer play.
Comments
The Game?
Lets understand something, MS did not create the numbers game -- their customers did. MS is just playing along. If the FOSS crowd does not understand that then woe is us. FOSS will always have a number foisted on it by the end customer -- always. It is how they make decisions.
That does not however mean that FOSS has to play the 'for-sale' game. But lets first throw away the 'free' canard. Any Fortune 5000 is going to be looking for support. The FOSS majors (Red Hat, MySQL, Postgres, etc) all provide professional support for a fee. For any corp, in their eyes therefore the software is not free even if they paid no purchase invoice.
So if its not free then what is the FOSS advantage? Lifecycle and Scale.
By Lifecycle I mean the ability to use the same software from prototype to datacenter without busting the budget. When you're building new ideas use FOSS. One of them pans our you still use FOSS. When the program gets to UAT the same FOSS is in use. Finally when it gets to production the licensed FOSS is substituted. All without the need to buy additional licenses for the early phases. The side benefit is that IT, if it has a standard suite its promoting, can not only accept the production code without complaint but support the early efforts without too much pain.
Scale of course in being able to utilize the same software regardless of the installation. FOSS has a better handle on that score. Nor does it play the fractured features game used by the proprietary purveyors. The added plus is that much of the scale issues for FOSS are pretty well known as well as their solutions.
The 'game' for FOSS is not to play the buy but the total corporate costs. FOSS should foster the adoption of IT FOSS standards. Promote the development of FOSS virtual appliances. Expand training opportunities at lower cost than the proprietary purveyors.
But lets not fool ourselves -- Free is not going to win the day.
Sounds like we agree
When I say numbers game I am not referring to price at all. I am referring to number of installs, number of sales, number of units and users. Whilst MS may not have created that game they do encourage its continuity. It's they who keep coming out with studies and reports all related to the fact that number of installations and best sellers apparently matter. Their customers have swallowed it hook line and sinker and whilst we need to understand that game we do not need to play it.
There is also a huge difference between companies who build their business around free software -- like Red Hat -- and the free software community. The former can and should play the numbers game as part of its business plan but -- because of the freedom in the software -- even they do it on different terms to Microsoft. The numbers for Red Hat are related to associated services not software licence sales. If the free software community plays the numbers game -- that is if it becomes part of our philosophy for creating software with freedom running through it then woe is us indeed.
You've mentioned the word free a few times in a context that is against what I meant. I do not consider cost an important part of software deployment, certainly not as important as freedom. Any software project -- be it a home office suite or a multi-seat business roll-out should always consider cost but not as much as the longevity and sustainability of the project and that is whee the freedom comes in.
When I say free software should not play the sales and marketing game, it is precisely because the cost is no longer the focus ( or at least it shouldn't be ). The cost still has a part to play but it is no longer the focus, the freedom is.
That is why and how the game has changed.
cheers Ryan
Why do we care about the numbers?
It's interesting to ask "why do we care about the numbers?"
I think if you look at it from the proprietary software perspective, the numbers are a way of seeing how "future proof" a piece of software is.
If it has a lot of sales and a strong company behind it, then it will continue to exist. And as long as it continues to exist, it will be supported (by someone, even if not by the originator) and your data (in proprietary formats) will continue to be accessible, and so on. (Among other advantages).
In other words, the customer is really looking for security.
Now, of course, free software solves these problems in much more direct and complete ways. FS communities are far more resilient to low deployment numbers than are proprietary companies, for example (because the future proofing comes from the free availability of the source code, not the continued survival of the organization keeping the secret). Similar arguments can be made in many other areas of user concern.
So, it can easily be the case that a piece of free software has greater "future proofing" and "data accessibility" than the competing proprietary software, even should that proprietary software control more than 99% of the market versus 1% for the free software.
In other words, the deployment numbers may simply not be relevant to the thing the customer or user is really trying to glean from them.
In fact, I suspect the performance of the FS packages's community is more likely a factor of its absolute size rather than its share of the market -- since the issue for FS development is "is the bazaar large enough to enable effective production?".
The question that arises, then, is how can you convert that information into a meaningful statistic rather than a qualitative hand-waving argument?