I personally can't wait until the freemium model finally dies. It is extraordinarily difficult to pull off despite how popular it is and creates desperate decisions like this.
Everything we offer is offered somewhere else for free, but we just package it all together in a much easier format. That's a tough sell out of the gate but once people get a taste of it, they understand why it's worth it to pay.
I've typically read that they don't work. For example: http://www.softwarebyrob.com/2010/08/18/why-free-plans-dont-...
All offer a non-trial/non-expiring 100% free home user version of their security software, with a more feature rich premium option available for payment.
Freemium works like gangbusters in some industries.
Also, you could make a case that these are not truly freemium, since any commercial use requires payment.
We have people routinely write in to us - 'Ive used your product for years and it's worked so great I just want to help contribute' and so they buy a license - just out of gratitude from years of free service and not because they are looking for any of the Premium features.
Regarding the 'cost of free users' - the cost is negligible. It's not like we write different definitions for the premium and the free - they all share the same definitions so there is not duplication of effort. Also, usually security software also acts as a 'data collector' in that it sends back suspicious files back to the lab for analysis. In this respect, having hundreds of thousands or millions of free users makes it easier to write definitions and protect all of our users because they provide us with more samples than we would of otherwise had access too. We actually ship a better product BECAUSE we have such a large # of users (both free and paying users).
Regarding 'cost of free users', I was thinking of support costs. Even though a user does not pay for the product, they still expect to get help when they run into trouble. And if they don't get help, that could wash away some of the goodwill they might otherwise feel.
Also, bandwidth costs for downloads.
Some companies do not offer phone support (especially not to Free users), so this limits free users support down to just tickets and forums (which can be user-moderated to a large degree). Some companies have had success bouncing free users who need support to 3rd party 'PC help' type companies who can work with the person over the phone - for a hefty fee of course, and part of that goes back to the originator as commission.
Dropbox has a freemium model and is probably profitable or on their way there.
One thing both have in common: they are basically giving away disk space. Storage costs are trivial -- we're approaching pennies per GB -- but giving away CPU time or bandwidth is a much riskier proposition.
I'm not saying freemium will never work, but let's face it: it's ubiquity is because it was at one time the flavor of the month startup model that VC money was pouring into, not because it is a superior or effective business model.
So the freemium model is like any other small business?
Not perfect correspondence, but pretty close. :)
1. Can you monetize free users with your product? In many cases you can, using ads, affiliate links, search partnership, ...
2. Is it cheap for you to support free users? This have to be compared to conversion rate. If it is really cheap to support free users than you can have bad conversion rate.
In some specific cases the freemium model is the only model that works: products where interaction with other users is required for product to be usable. Only example I can think of right now are online games. You need players in order to attract more players, some of which may convert to be paying customers.
* Charge for your game
* Build up a loyal fanbase
* Continue to provide actual new content for players, for free.
* Start selling hats
* Stop charging for the game (nearly four years later).
I wouldn't necessarily mark up TF2 as a freemium success as it only went to the model fairly late in it's life (after it had already more than made back it's budget, likely).
You can get many users to sign up for free and they can be great advertisers for you.
However, providing something for free can also trick you in deciding which features to build and whether people really want the thing you build. The difference between free and $1 is much bigger then $1 and $10. I believe that is where a lot fremium startups fail. They just assume 5% of free will convert. Nope: in many cases it is 0%.
So I never sign up for service which is free and there is no way to upgrade to some paid version. Even if they are in beta.