Why are big software companies going freemium?

There's a lot of criticism of freemium in the startup community, but in 2009, some of my favourite large software companies moved towards, rather than away from it.

Atlassian started their $10 charity starter packs of confluence and Jira.

Jetbrains offered their flagship IDE as a free community edition.

Fog Creek created their free student and startup edition.

IBM are also making $50k Rational tools free for small development teams.

I'm sure there are others.

These are all companies that previously did not operate freemium. More tellingly, in the Atlassian and Jetbrains examples, they took their core product, and basically gave it away for free to millions of small businesses.

What do you think this movement occured? Bad economy or great way of spreading the word and getting in at the grass roots of startups?

As a startup targetting small businesses, should this concern me? Maybe this trend is based on the fact that micro and small businesses that are emerging are just not buying software? Maybe it's an argument to be looking at larger businesses and the enterprise?

One reason is to offer the typical "The first one is free" drug dealer concept. Once you are hooked you need it. Then when you grow you can afford it and will pay for it.

They don't cannibalize their sales since the people taking advantage of the free versions wouldn't have bought it, but they also eat away at market share of competitors.

You missed a biggie -

Microsoft launched bizspark - basically "Free" msdn for small/new companies. My guess is that they were losing too many young/startups types to linux, etc. Note also that they get HUGE amounts of data and first looks at small businesses that may be targets for acquisition.

Marketshare/mindshare. That is my vote for why many of these players do it.

EDIT

In response to your comment about jetbrains - they have many products - not just an IDE. This was likely (just a guess here of course) because there are other IDEs out there that were really good and perhaps causing them to lose focus on their other revenue generating products. If it costs too much to keep up with the competition and you have other profit leaders, then it might make sense to drop the cost and do other things. (again, that is speculation on my part)

I think there are some important aspects to the Fremium business model that exist in the examples you cited. I'm not sure if this is the case for all companies with a freemium model.

1. Switching Costs - In some of your examples (FogBugz & Atlassian), there is a network effect in place. These tools are targeted at making teams more effective, and as the team grows, it will want to use the same tools it's used to, so there is a switching cost to move to a new tool. You don't want to make all the existing team members learn a new tool just because you added one more person to the team. These companies are making the bet that while every team won't grow over time, enough will to make it worth it. I imagine that it's harder to make freemium work with a product targeted at individuals.
2. Low Marginal Costs - The cost for Atlassian to provide their software to a small team rounds to zero - it's basically the cost of the bandwidth for the download. This means that it doesn't increase expenses for them to offer the product for free. It's harder to make freemium work if you have to incur a bunch of costs with each free customer.
3. Healthy Base of Paying Customers - It's a lot easier to give your product away for free to some custoemrs if you have lots of other customers paying the bills. That's the case for all of the companies you cite, and it allows them to take a long view (these people might be paying customers in a few years) because they don't have to worry as much about today's revenue. I think it would be harder to start a brand new company using the freemium model unless you have adequate initial capital.
4. Not Cannibalizing Existing Revenue - In the cases you cite, the small teams probably wouldn't use the product if they had to buy it. There are enough free options out there that, while not as good, would work just fine. This means that there is no revenue lost by offering the free option. It would be harder to offer Free if it meant that a percentage of your paying customers would move to the free option.

When budgets are under stress companies aren't going to try out new, expensive software, to see if it is better.

By basically letting some people use it for free it will help show whether the software is useful or not.

Since, in 2009, we saw many companies have to start doing some belt-tightening, this may be why there was an increase then, as, in the past, these companies may have been getting enough sales that they didn't mind if they missed some opportunities, but now they need to make certain to try to get as many customers as possible.

If your product has a nice split, and you can make a compelling argument for the pay version, then the freemium idea has value, but this model has many risks associated with it, as well as possible benefits.

In the case of jetbrains, there is another factor to weight in: Today's IDEs heavily rely on plugins. IntelliJ competes with Eclipse and Netbeans, both of which are known for their abundance of Plugins (especially Eclipse). The free version of IntelliJ might attract plugin makers since the potential potential market share is much larger than without a free version.