I recently read Kane Bennett's post, "Make it Free or Fail".  I agree with most of Kane's conclusions, but I actually think he frames the problem incorrectly.  I think he wraps two different trends under one, the first is free software that is ad / sponsor / API revenue supported.  The second is a freemium business model.  These are completely different trends and while they both result is one releasing their software for free the goals are completely different.

 

In the first trend, pioneered by Google (they weren't the first, just the first to do it really really really well), and eventually leveraged by a number of other startups such as Facebook and Twitter was the theory that you attract eyeballs with a great product that is free to use, but supported by a mixture of ad revenue, sponsorship dollars, or taxing use of an API/app ecosystem (Google / Twitter).  

 

The second trend is the freemium business model.  According to Wired, Free! Why $0.00 Is the Future of Business, the freemium business model was pioneered in 1895 by Gillete.  In software, the freemium model usually manifests itself as a free feature-less version of the software and a paid feature-rich version of the same software.  The most popular early example of this that I can think of was Flickr, but it has been used successfully by Dropbox and Skype (as noted by Kane). 

 

There is also an interesting third redux of these ideas in Pandora.  Pandora isn’t really freemium or ad-supported; it is both.  There is an ad-supported product that is free, and the same product available ad-free for a fee.  I wont go into this trend at all, but it is worth noting.

 

Kane is abandoning the “free software movement” as he calls it because he believes it has been broken due to too many people participating in it.  This hypothesis relies on the assumption that the reason free software is more likely to go viral than paid software is because there are fewer free software options out there and thus it is easier to differentiate your product by making it free.  Therefore the conclusion follows that if there are LOTS of free software options it is no longer an important differentiator and thus you will still need to spend money on sales and marketing which (according to Kane) breaks the free software business model. 

 

This is absolutely NOT why free software is more likely to go viral than paid software.  In Predictably Irrational (a MUST read if you haven’t), Dan Ariely explains the why free is “less than 1 penny less than 1 cent.”  He proves through a variety of experiments that free is in fact a special price all its own – it is not just 1 penny less than 1 cent.  This explains why free software “can” (not will) have better uptake with less marketing than paid software.

 

Should all Software be Free?

Absolutely NOT!  Free software is a tool – a great tool – but a tool, no more no less.  It is certainly NOT magic pixie dust to be sprinkled on every bad idea.  Here are four great reasons to release software for free:

 

Freemium Business Model

If your business model is based on the idea that some portion of your users will pay for advanced features while the rest use the free version, then a free version of your software is a great tool to help market your platform and give people something to test out – Predictably Irrational establishes that people really are more likely to try something for free.  This assumes that you want EVERYONE to try your product.  A freemium business model is not for everyone, however.  For example, if your product is aimed at a niche market, or the cost to install / switch to your software is high then you can’t really make the “cost” of your software free and thus it is virtually impossible to truly have a freemium business model (and a bad idea to release a free version of your software).

 

Ad-Supported Business Model

In this day and age, this is almost tantamount to saying your software should be free. For one thing if the main revenue potential in your product comes from eyeballs then you should do everything possible to make your product easy to use – one of which involves making it free.  Again, as stated in Predictably Irrational, this is the cheapest way to make it interesting to try out if you are a consumer.

 

Platform-Play

This is not nearly as well established a reason as the two above.  There is currently a test going on in the marketplace for this: Windows Phone 7 OS vs Android.  Though one is free and one is not, the end phones are never really free (or atleast Microsoft and Google have very little control via the price of the OS whether the phone is offered for free or not).  So to the consumer an Android phone might be cheaper (though I am not sure that is true) but it wont be free. This WILL NOT leverage the value of “free” in the consumer’s eyes.  Thus this test is going to help us figure out whether phone makers are impacted by the price of “free” in the same way consumers are or not.  Phone platforms aside, a platform is still a great place to release software for free.  If your main value is users (similar to ad-supported businesses) then you should focus on the cheapest way to get them and a free platform is probably one very good way to do that.  This doesn’t mean that a free platform WILL go viral or even have better uptake than a paid one, but a good AND free platform will always outcompete a good AND paid platform.

 

This is a damn good idea… and some day we’ll figure out how to make money

In Paul Graham’s, The 18 Mistakes That Kill Startups, he postulates that it is much harder to make something people want than it is to make money from something people want and further that very things that are truly worth something to their users cannot be monetized.  Obsiously it is a risk that you wont be able to monetize it, but it is a much bigger risk that you simply wont be able to make something people truly want.  If you are a startup with a great idea, but not one you are sure you can make money off of, it is probably OK to consider that a risk that you will address in a later version of the product and instead focus version 1 on the core risk – is your theory that this is a quick ass idea (and that you will be able to execute on that idea) correct?

 

When Software Should not be Free

Kane mentions that every angel investor says, “just release your software get a million users and then get 1% to pay for it” but few ideas spread that way and thus releasing the software for free is a misnomer, and can actually detract from figuring out the answer to your core question, will people pay for this?

 

If your core risk is “will people pay for this”, then Kane is absolutely right don’t release your software for free, release it as a paid product first.   At the end of the day entrepreneurship is all about de-risking a very risky idea.  That requires you, the entrepreneur, to find creative ways to validate your hypotheses very inexpensively.  To do that, it is all about leveraging every tool you have to accomplish this.  One of those tools is free software.  Hopefully, by now you believe me that free software will spread more easily than paid software.  It can therefore help you address what is usually the most pressing risk – have I actually created wealth?  Have I created something people really want?  If your core market is a much smaller niche market, or if you are more concerned about answering the question “will people pay for this” then free software is probably not going to help you address any of your core concerns – therefore it simply isn’t the tool for the job.

 

A great example of this is Gmail.  They created a new email application at a time when other email applications (both cloud-based solutions and desktop solutions) were quite sophisticated.  Gmail’s core thesis was that, conversation based email, a never delete anything amount of storage, and a search not browse philosophy were the core of a killer email product.  They never could have charged for a product that was missing common features like tasks and calendar integration.  It would have cost them a substantial amount to build those features into their product.  Instead, they released the product for free enabling them to test whether their hypothesis about those features was correct or not – as history has shown, they were indeed key features.

 

Magic Pixie Dust Doesn’t Exist

Magic pixie dust doesn’t exist.  Nothing can turn a bad idea into a good one.  Since this is really all about testing, the other important thing to note is that it is also key to have the right expectations.  Whether you release your software paid or for free, you need answer to the question, “are we on the right track” as quickly as possible.  This means you need data on whether users actually like your product.  This requires having a hypothesis (that includes measurable expectations) and then testing those hypotheses.  If you release your product for free with the expectation that it will go viral and produce a million users in the first 3 months and it doesn’t then it failed your test. 

 

The key here is having the right hypothesis.  Does your business model require the kind of growth that comes from being viral?  If so, then you better test that your product can go viral.  Does your business model actually rely more on your users (however large or small) using it more hours out of the day?  Then don’t pay attention to number of users right out of the gate, pay attention to "time spent using application" or some related stat.  The point here is identifying key risks and designing experiments that prove to you that you have mitigated those risks.