On the other hand, most larger proprietary software businesses devote a surprisingly small amount of their operating budget to development. In my experience it's usually around 10%, but it depends on the company. I certainly haven't heard of anyone spending more than about 30% and there are quite a few big names that spend in the low single digits.
Potentially, if you can leverage the benefits of free software to pay for some of the things you would usually have to manage yourself, then perhaps you can reach some kind of parity even if the vast majority of users do not pay money for the software. Mozilla currently pulls in relatively huge amounts of cash (hundreds of millions per year), but when they first started it was considerably smaller. Because they were able to concentrate on development rather than marketing and sales, they were able to get quite a lot done with that money. Where they did spend on marketing (for example the famous New York Times ad), they were quite careful about it.
There are other successful free software plays. For me the biggest example is Red Hat. When these kinds of conversations appear, I always point people toward Michael Tiemann's chapter in "Open Sources: Voices From the Open Source Revolution" [0]. Cygnus software went from $6K to being sold to Red Hat for $600M, and Red Hat went on to great success with essentially the same business model.
My favourite quote from that chapter, talking about discussing his ideas for a free software business model: "I never got farther than 'It's a great idea, but . . .,' when I had my second insight: if everybody thinks it's a great idea, it probably is, and if nobody thinks it will work, I'll have no competition!" Even now, I think there is virtually no competition in Red Hat's space because nobody thinks their business model will work.
In the end, I think that developing free software business models is where it has to go. You can't charge for the software. I personally feel that dual licensing is not really a free software business model because you don't actually make money from free software -- you are just using it as a loss leader for your traditional proprietary software. Like you imply, in the enterprise space you can probably do well (for example, by copying Red Hat, or by making a foundation like Mozilla or Apache). But there is a lot of other software out there. I think in those spaces you have to be a lot more creative about what you charge money for. People will clearly spend money -- just look at the horrible, horrible "free to play" game market. The trick is to allow them to do so in a non-horrible way ;-)
> On the other hand, most larger proprietary software businesses devote a surprisingly small amount of their operating budget to development. In my experience it's usually around 10%, but it depends on the company. I certainly haven't heard of anyone spending more than about 30% and there are quite a few big names that spend in the low single digits.
Now I'm curious; what else do they spend money on? Sales and marketing of course, but what else?
Generally speaking it doesn't include management (except for immediate dev management). It also doesn't include support or operations (in a software company that would include secretaries, security, cleaning staff, etc). It usually doesn't include capital expenses either, so depending on how you do the accounting, things like running the building is not included. From that perspective, the numbers can be a bit deceiving.
Depending on the area, the costs can swing wildly. I worked at Corel back when shrink wrapped software was a thing. Sales channels cost huge amounts of money. Even after that period getting your software into OEM builds cost a fortune. That's stuck in "costs of sales", usually. I've worked at companies where costs of sales was well over 50% of revenue. As you can imagine, there is considerable latitude for corruption ;-).
If you are running a service, "operations" often pays for compute infrastructure, but it depends entirely on the company. You might even have dev ops in operations so that can be misleading because you don't have a product without the service.
My experience has been that management, operations and sales takes 60-70% of revenue. The absolutely necessary expenses in operations usually amount to less than 10%. Capital expenses can be very large if they are investing in real estate or something similar, so that's a complete wild card. Sales and marketing are kind of linked because usually you have to spend more on one if you spend less on the other. How much you can get away with if you change the business model is kind of up in the air and I'm really speculating when I suggest that a free software model will be cheaper -- I believe it, but obviously have zero data.
It's worth downloading quarterly/year end reports from various high tech companies and reading them. It's pretty incredible how much information you can get from them. They won't break down everything exactly the way you want, but you can usually figure it out -- and often see pretty plainly where they are playing games.
Complete diversion: One of Corel's favourite tricks in the past (though not now, I think -- everything changed after they were sold) was to overfill the sales channel in first quarter. Basically they would "sell" the software to dealers, but would provide "refunds" for software that wasn't sold by the end of 4th quarter. So 1st quarter was always massively positive, while they were always massively (and somehow unexpectedly) in the red by 4th quarter. Executive stock options were granted 4th quarter and there was a trading window in second quarter. So the stock would plunge dramatically (one or two binary orders of magnitude) just before the stock options were granted and then rocket back up in time to sell them. They did that year after year after year (I was there for 5 years and it happened every time). The reason for the swings was literally written in the quarterly reports. In first quarter they would say "We sold $X, but this is subject to 4th quarter refunds". Then in fourth quarter they would say, "We experienced a one time loss from refunds in the sales channel". I swear nobody reads those quarterly reports... :-P
> Complete diversion: One of Corel's favourite tricks in the past (though not now, I think -- everything changed after they were sold) was to overfill the sales channel in first quarter. Basically they would "sell" the software to dealers, but would provide "refunds" for software that wasn't sold by the end of 4th quarter. So 1st quarter was always massively positive, while they were always massively (and somehow unexpectedly) in the red by 4th quarter.
More or less OT, but maybe interesting: This was the strategy that killed West End Games (once one of the biggest publishers of pen & paper games). They would print as many books as they could and send them out to shops, but the shops had the option of sending them back when they couldn't sell them. Then in 1998 for some reason almost all of the books came back and WEG was bankrupt. Seems it worked far better for Corel. Not sure if this is good or bad.
Corel was actually very close to bankruptcy at that point. I think they were down to $2M and Microsoft bailed them out by buying a new round of preferred shares. If I remember correctly it was $130M. The CEO of Corel at the time got quite cosy with MS and decided that they should be friends -- with Adobe being the "enemy". We seriously had company wide meetings where they repeated "Kill Adobe" over and over again.
The strategy that MS suggested Corel follow was to invest the $130M in acquiring companies (which is how they ended up with Paint). Now here's the fun part: MS's preferred shares were non-voting, but they had a veto on acquisitions ("Just so that we can protect our investment. Trust us, we'll never use it! We're the ones telling you to acquire companies, aren't we?"). MS, then turned around and sold their shares to a VC company called Vector (part owned by Paul Allen).
After buying the shares, Vector said, "We're putting a veto on all of those acquisitions. Oh... the penalty clauses will make you bankrupt? So sad for you... We're happy to buy your entire company and turn it private. If you don't agree, then I guess you are out of business."
So Vector managed to buy up Corel for (IIRC) somewhere around $110M. At the time Corel still had something like $90M in the bank, so Vector paid $20M plus the $13 million form their initial purchase from MS. They managed to do the acquisitions and then resell them. After a year, they made a new public offering of shares for 25% of the company for somewhere around $40M (these numbers may be wrong as they are coming from my aging memory, but they are illustrative of what went down if not accurate) -- giving them a pretty awesome ROI in just one year.
All of the senior Corel executives (including the CEO and VPs) somehow managed to land jobs as "senior product managers" at MS. Some of them seemed to turn down their lucrative parachutes from Corel to accept the job, so... I guess they were given handsome signing bonuses...
It shouldn't surprise me, but once you get above a certain level it stops being about making a product or doing a service and becomes all about making money. The companies are just a vehicle for that pursuit and in a lot of ways I think the "power people" don't care that much if the company survives or dies. They just care where the money falls.
Potentially, if you can leverage the benefits of free software to pay for some of the things you would usually have to manage yourself, then perhaps you can reach some kind of parity even if the vast majority of users do not pay money for the software. Mozilla currently pulls in relatively huge amounts of cash (hundreds of millions per year), but when they first started it was considerably smaller. Because they were able to concentrate on development rather than marketing and sales, they were able to get quite a lot done with that money. Where they did spend on marketing (for example the famous New York Times ad), they were quite careful about it.
There are other successful free software plays. For me the biggest example is Red Hat. When these kinds of conversations appear, I always point people toward Michael Tiemann's chapter in "Open Sources: Voices From the Open Source Revolution" [0]. Cygnus software went from $6K to being sold to Red Hat for $600M, and Red Hat went on to great success with essentially the same business model.
My favourite quote from that chapter, talking about discussing his ideas for a free software business model: "I never got farther than 'It's a great idea, but . . .,' when I had my second insight: if everybody thinks it's a great idea, it probably is, and if nobody thinks it will work, I'll have no competition!" Even now, I think there is virtually no competition in Red Hat's space because nobody thinks their business model will work.
In the end, I think that developing free software business models is where it has to go. You can't charge for the software. I personally feel that dual licensing is not really a free software business model because you don't actually make money from free software -- you are just using it as a loss leader for your traditional proprietary software. Like you imply, in the enterprise space you can probably do well (for example, by copying Red Hat, or by making a foundation like Mozilla or Apache). But there is a lot of other software out there. I think in those spaces you have to be a lot more creative about what you charge money for. People will clearly spend money -- just look at the horrible, horrible "free to play" game market. The trick is to allow them to do so in a non-horrible way ;-)
[0] - http://www.oreilly.com/openbook/opensources/book/tiemans.htm...