Best Presentation on Software Business and OSS

stephe writes "Brent Williams presented 'Open Source Business Models: A Wall Street Look at a Wild 2006 and the Prospects for Even More Fun in 2007' at EclipseCon last Tuesday. Brent is (temporarily) an independent equity research analyst, who moved to Wall Street after 20 years in the software trenches. He starts with a tear-down of the Oracle Linux debate and the Microsoft Novell deal. I especially like his taking apart the commoditization myth and his observations around interface standards versus standards of implementation. He graciously allowed me to post the slides on my blog. They're getting a lot of interest from the open source business crowd, and I thought the Slashdot crowd would want to see them as well. Enjoy."

Comments

premium brands ignoring price competition?

[turing_m]

FTA:
"What price changes did Red Hat make immediately in the wake of
the Oracle announcement?
- None. Zero, zip, nada.
- We're not hearing of any individual deal discounting.
- Red Hat knows that they have a premium brand, so ignoring people
competing on price is the right strategy.The role of a premium brand
- Lamborghini ignores price competition between Hyundai and Kia.
- Oracle ignores price competition between MySQL and PostgreSQL."

This is not the case at all. In the last few years, MySQL has matured and more people have found out about PostgreSQL (in fact, PostgreSQL is probably the best kept secret OSS has to offer - it has a kick ass feature set and it's completely and utterly free). For a large amount of enterprise stuff, PostgreSQL is more than adequate and as a bonus, does not treat your data as garbage.

Anyone considering building some sort of database application has the option of spending a couple months (with change left over) from the money they would spend on an Oracle license, and invest it in learning PostgreSQL. At the current rate of developement, it will in all likelihood solve any future problem they could have. For free. No worries about licenses. Anyone in a startup where money is tight and time is cheap should be considering PostgreSQL.

This has had direct ramifications on the strategy of all the big database players. At the very least, they all now have a free entry level option to compete with OSS competitors.

Re:premium brands ignoring price competition?

[turing_m]

"They say that if you make a RFQ and get one offer for $100k, three for $10k and one for $1k, they're likely to drop the $100k (too expensive) and the $1k (must have missed something). MySQL and PostgreSQL score about 10x as high on the WTF scale to most PHBs. If anything is free* in their world, they expect a bait-and-switch like *upgrade now to $foo pro for the good features, that was just the hook."

I suppose if you are dealing with a PHB you have your work cut out for you, and in all likelihood at this stage of the game your boss will side with one of the major players.

A major niche (and it's really a wedge, IMO) of PostgreSQL is in the area of heavy cost constraint, used by a smart individual with some time up his sleeve. This could either be a start up business (e.g. run by a college student/grad) or in a small-medium sized business where a technically savvy employee wants to make a name for himself by building some sort of money making/saving database application and has enough time to do so. It could easily happen with a (perhaps fairly computer illiterate) founder who pinches pennies to the extent that he might begrudge even the $1000 option. Often founders of companies are this way - pinching pennies was how they got their start.

Another market niche is the fairly low priced database application needing a decent backend. You can undercut your competition knowing full well your application will work, and still make more money.

A few of these examples, especially successful startups, and PostgreSQL will build the sort of momentum that linux built 5-10 years ago in the server market. I think it is already happening. What will slow it is the difficulty in transferring from an existing database installation to an OSS version. Switching firewalls, file servers etc. is not as nearly as troublesome as database migration (from and to any platform).

However, the database functionality is there. PostgreSQL holds your data well. It hews close to the SQL standards. Complex joins are no problem. With a little work, you can make your backend sing with functions, triggers etc. And it's fast, when you are actually comparing apples to apples (i.e. you are comparing two ACID compliant RDBMS' with referential integrity constraints, not a mythical database consisting of MyISAM performance with InnoDB functionality).

I suspect the rise of PostgreSQL will be more meteoric than most pundits are expecting.

Re:First thought from first slide...

[Zapraki]

Ya, sorry, I didn't mean to sound snobby. It just frustrates me when people insist on sticking with what is an inferior solution (imho), when much better ones are there for the taking, with basically nothing extra required.

Although I guess the very weakness of hotmail is what could prevent people from changing. No forwarding?? If somebody had a long list of contacts who emailed that account, they might have to stick it out rather than risk losing contacts with a switch over. Or else maintain multiple accounts, which could be even worse.

And yes, haha, may Outlook burn and burn.

Well he makes a lot of errors

[SmallFurryCreature]

For instance claiming that there are a large number of producers and consumers in the commodity market.Eh what? How many commodity producers are there? Flower mills, power generators, oil producers etc etc. Not all that many and they are merging all the time.

He also claims that it is easy to go into that market, yeah right. Isn't it exactly other way around usually, hard to get into a commodity market?

I think we got the wonders of reverse logic in a powerpoint presentation at work here.

The trick is to put your conclusion on a slide and then keep inserting pages in front of it to make sure that you arrive there.

Yes his end conclusion makes some sense, that software is not a commodity and that opensource business model can work, BUT the logic he uses to get there is seriously flawed.

The simple fact is that economics are a lot more complex then he makes them out to be.

Take the product water, getting into the tap water market is next to impossible. Just try it, go ahead, dig up the whole town to install new pipes, find a clean source of water, put installations in place to get it to drinking level and setup up the whole supporting operation. No hope.

Getting into the bottled water market is a lot easier, just find a clean spring and get a cheapo bottle plant and voila, new brand product and watch the millions roll in.

So is water a commodity? No, not by this guys definition, tap water don't count because there is only one supplier and bottled water don't count because it is not massive enough. Brand loyalty is extremely low. How many people do you know that drink bottled water who won't drink tap water (in countries were tap water is drinkable) and remember, the water you make coffee with counts.

But his greatest mistake is this. Red Hat does NOT sell software. It sells support. Read this and understand this. Red Hat is NOT in the production industry it is in the services industry.

Go back a few pages and read his list of products in commodity markets, he mentions wheat, pork bellies, cars, chips nails etc etc.

ALL physical products.

He doesn't mention banking, insurance, support, even labor services. Red Hat doesn't have a physical product, it doesn't BUY a ton of plastic to produce 10.000 RHEL discs. It sells a service, the service of support a software package that you can just as easily get for free.

The service industry is a WHOLE different beast, for instance it is almost always far easier to get into. All you need is a phone line to start a support line after all. Even banking and insurance are easy, all you need is a bit of cash. Loan sharking is excellent proof of just how easy it is to get into banking (easy ain't the same as legal).

The service industry is also different from the production industry in another extremely simple way. Labor.

If Vista had been a physical product MS could have easily finished it on time by just hiring more people, if you want to build a road faster you just have more people working at it, more or less you can just keep upping the labor force to increase production.

Service industry on the other hand can't be automated. You got production plants that are run by one old man while next door the office has hundreds of people manning the phone lines. Luckily scaling this up is easy and relativly cheap. It don't cost that much to install another 100 phones lines for support.

Anyway, I think the article is just plain wrong in its basic logic. THe conclusion might be correct but that is because the author already arrived at that beforehand.