OnStartups

Startup Pricing Models: Free Forever, Freemium and Freedom To Pay

Posted by Dharmesh Shah on May 15, 2006 20 Comments


Of the articles I’ve written on this site, the one on giving your software away for free continues to be one of the most popular (currently ranked as #2 on the site).  In response to that article, I’ve received a number of comments.  Some highly insightful, some a little misguided and some a result from what is likely misunderstanding.

I’d like to take one more crack at this whole “free vs. not free” aspect of software pricing for startups.

Free vs. Not So Free For Software Startups
  1. Free Forever:  This would probably better labeled as “subsidized”.  Basically, instead of charging anything directly to the customer for your software, you’re generally charging someone else.  The best and most obvious example is advertising.  The second, and significantly more subtle example is when you are actually charging your future acquirer for your software.  In this latter model, you amass users, let them use your software (and likely your bandwidth, storage, etc.) in the hopes that somebody will acquire you and pay for those users some day.  I’ve already beaten to death, in prior articles that advertising revenues are sketchy at best except for the biggest players – and even then, the predictability of this revenue stream is suspect.  And, if you’re expecting your future acquirer to ultimately pay, then that’s risky too.  Have already written a lot about this, not going to repeat it here.

  1. Free To Use, Pay For Service:  This is a common practice with some of they hybrid open source companies out there.  Basically, the product is given away for free and the customers are charged for training, customization or other value-added services.  I don’t have anything against this model.  But, profit margins are usually lower because it’s essentially a service business instead of a product business.

 
  1. Freemium:  I think Fred Wilson of Union Square Ventures originally coined this term,  and it’s useful (much more than “Web 2.0”).  With this model, you basically give away a free (as in beer) version of your product with the hopes of upgrading some number of these free users to your “premium” version for which you charge something.  The idea here is that the cost you incur for supporting the free users is a marketing cost, much like other marketing costs.  And, much like other marketing costs, this is ultimately borne by your paying customers.  So, in a way, your paying customers in the freemium model are subsidizing your free customers.  But, customers pay for your advertising and other distribution costs anyway (it’s embedded in their price), so there’s nothing really new here.  Note once again, this is a marketing path, and that your real (as in paying) customers fall into the following category.

  1. Freedom To Pay:  This is the classic “customers pay for value” model.  I’m turning this around a wee bit and suggesting that with this model, customers have the freedom to pay.  Why, you may ask, am I positioning paying money as a benefit to customers?  Don’t customers want free?  The answer is, not always.  By giving customers an opportunity to pay, you are also giving them the freedom to expect some value from you.  Though customer expectations can (and do) exist for “free forever” products, it’s rarely the same as that for a paid product.    If you doubt this, try extracting a meaningful “service level agreement” (or for that matter, any agreement) out of a company giving you software for free.  In the cases I’ve seen, the companies are expecting something from you (like waiving their liability) and not the other way around.


Google API Example:  One (somewhat unrelated) example is the Google API.  There are likely better examples, but I’m being burned by this one as we speak, so it’s on my mind.  Google provides an API that developers can use to access Google services (such as search) through a programmatic interface.  This is cool.  Access to the API is free.  That’s cool too.  What is decidedly not cool is that there is an upper limit to the number of queries I can do in a day with my Google account (it happens to be 1,000 right now) and I have no way to increase this limit.  To be clear:  Google will not accept my money in order to increase the number from 1,000 queries to 10,000 (or 100,000) queries.  Basically, I don’t have the freedom to pay.  In this case, I’m clearly not a customer at all, I’m a beneficiary.  In some cases, this is fine.  In many cases, it’s not.  I’ll bet there are a number of customers that would like to have the freedom to pay in exchange for the ability to expect value.  On a related note, the Google API is still in “beta” so not only do I not have the freedom to pay, I don’t have the freedom to even know if I’ll ever have the freedom to pay as Google won’t talk about pricing (if there will be any) at all.

A few additional notes:
  1. If you’re playing the “freemium” game, you need to look at this for what it is:  a marketing expense.  If it takes 100 free users to get 1 paying customer, you may have a problem (because the 100 free users are costing you something, and that’s contributing to the acquisition cost of the customer you got).  My guess is that conversation rates vary widely, but let’s use 1% because it makes the math easy.  If it costs you about $2/year to support a customer, that means you have $200/expenses for your 100 free customers that it took to get you one paying customer.  Maybe, with your pricing, this makes sense. Maybe it doesn’t.

  1. I have nothing against free trials of software products.  I think this is a great idea (and almost a market requirement these days).  

  1. There’s no such thing as a “zero cost” user – even in software.  Though hardware, bandwidth and storage are cheap (and getting cheaper), the costs are still greater than zero – and since there are atoms involved, and not just bits, they likely will never quite be zero.  [Author’s note:  I use the term “atoms” to loosely refer to physical things like computers and infrastructure.]  So, if you have 100,000 free users, they’re still not free of cost – just possibly free of revenue.

  1. Please spare me examples like MySpace, Facebook and others.  Companies in the “social networking” space are basically in a race for critical mass.  Some will make it to the finish line, most won’t.  If you want to run in this race, more power to you.  May the wind be in your sails…


Summary Of My Point:  Don’t confuse marketing models with business models.  The former helps you get visibility for your product, the latter defines how you will actually make money.  Too many startups try to pass off marketing models as business models.  Also, make sure you think through what your customers likely want (you know, customers, those people that give you cash).