The freemium business model: giving away pays

I've recently been thinking a lot about the freemium business model. For those unfamiliar with the freemium business model, it was first articulated by venture capitalist Fred Wilson in 2006:

"Give your service away for free, possibly ad supported but maybe not, acquire a lot of customers very efficiently through word of mouth, referral networks, organic search marketing, etc., then offer premium priced value added services or an enhanced version of your service to your customer base."

I've been thinking about the freemium business model because, inspired by Drupal and Open Source, both my companies, Acquia and Mollom, use a freemium business model. (Technically, Acquia uses an Open Source business model which is different from the freemium business model, but there is plenty of overlap and similarities -- pointing out the differences could be a blog post and discussion on its own.)

At Acquia, we currently provide community subscriptions for free -- people that want help with Drupal installation and configuration can get free support from Acquia's Drupal experts. While our free support is limited to certain channels (i.e., forum only), certain support questions (i.e., no module development help and no security best practices) and comes without response time guarantees, we have people on staff whose full-time job is to help you (example customer story). Further, we invest heavily in Drupal and give those contributions away for free.

Similarly, at Mollom, our basic spam filtering service is available for free to sites with limited post volumes. Our free website protection service provides all the features of our commercial Mollom Plus product, but is limited in the number of posts it will protect each day and in its access to our high-availability back-end infrastructure. The great majority of our Mollom clients are using our free filtering service with great success.

There are a number of things that attract me to the freemium business model. The first, and certainly foremost, is the opportunity to do “good” and “well” at the same time. It’s a great thing to help people build quality websites with Drupal, and it’s a great thing to provide Mollom to help deal with spam. Second, I believe a company is better off with a large install base than a small install base, even if the majority of clients ride free. A large install base translates to direct and indirect network effects, including efficient marketing, greater brand awareness, the collective intelligence of your users, and faster product adoption. And, last, I strongly believe that a successful company built on the freemium business model is simply a stronger and more defensible business in the long run.

The freemium business model is relatively new because it didn't become a serious option until the internet gave us a low-cost distribution channel. Ultimately, I can't help but think the freemium business model is the business model of the future for the sole reason that it puts the customer first. With the freemium business model customers only have to pay when they get significant value from the software (i.e. they have reached the limits of the free version). Compare this to the current model where people have to pay to get access to the bits, or where people have to pay before they got enough value from the software (e.g. most shareware software).

That all sounds great but you have to make the freemium business model work first. Getting free users to convert to paying customers is hard. Conversion rates of less than 1% are not uncommon. Free is often “good enough” and only a few people choose to pay for additional features and services. You have to put enough value in the free version to drive adoption (so that you get the scale and the network effects that derive from it), while providing enough incentive for people to pay for premium features or services. The marketing and sales funnel is really wide at the top, and very narrow at the bottom. Plus, you have to make sure that the paying users subsidize all the free users.

Achieving the right balance between free and paid customers is difficult and requires close attention to a number of variables. As a result, I've been trying to answer questions like: how much should we invest to acquire additional free users? How do you estimate the value of a free user? What is the cost of a free user? How long does it take for a free user to convert to a paying customer, and how many will do so? What are the triggers that convince free users to convert?

For example, in Mollom's case, one could argue that we get thousands of dollars worth of value from free users already. We currently have more than 3,000 active users that use Mollom for free. Say each user spends on average 15 minutes a week moderating his site's content and reporting classification errors to Mollom. Mollom learns from this feedback and automatically adjusts its spam filters so that all other Mollom users benefit from it. At a rate of $10 USD/hour, we get $390,000 USD worth of value from free users a year -- 3,000 users x 15 minutes/week x 52 weeks/year x 10 USD/hour = $390,000 USD/year. If these numbers hold up, the value of a free Mollom user could be estimated at $130 USD/year. And that doesn't include the marketing value they add. That said, the value of a free user probably declines as you get more of them and the business becomes stronger.

Both Acquia and Mollom have just opened for business so we have a ton to learn. It will be interesting to look at the different variables and questions a year from now, and to see what we have learned. I hope we can make it work so we can do good and well at the same time ...


bertboerland (not verified):

One can make free^H^H^H^Hcheap jokes like "the first shot is free" or "free, but we make it up in volume". Both are actually true!

The truth is that for both cases "free" is a very good business model. Let's first look at "the first shot is free" model.

I was only hooked to the Flickr service, once I was able to post from my iPhone towards my flickr account. And once I was hooked, in no time I was above the 300 pics a free user can have. Now comes the beauty of Flickr, you have the last 300 pics sliding in your free account and can activate older pics to be seen again in your stream by paying! A really great "first shot is free" business model with some vendor/data locking that might be debatable.

If you look at Mollom, the paid service would not be possible if the free service would not be there; sites using the "free" version contribute something. They give input for what is spam, ham, false positives and negatives and /create/ thereby a value that can be sold to others (with some additional services like uptime, max repair time etc). In this scenario ("free but we make it up in volume") the service provider gives away a service for free but gets more data back, aggregates this, adds value to it, combines it with other services and markets it as an commercial service. It also provides an upgrade path for the user, if the service becomes so important that your business relies on it, you /can/ pay for a better / faster service.

So both models have proven to be very good for the business (and for the user)!

Jeff Whatcott (not verified):

Good topic, as always.

Freemium is all about creating flow. If you can get a flow going, you can usually find a way to monetize. If you don't have flow, nothing else matters. It sometimes takes many tries to get monetization right, but you can't shape a flow you don't have.

One of the most beneficial flows that you get from a freemium model is the flow of feedback and market requirements. When you have a lot of eyes on something, you get more feedback sooner, and that allows each product iteration to focus on more important things early on in the iteration process. This is incredibly valuable to early stage companies because it lowers the stakes of product development decisions. It gets things out of the opinion domain and into the facts domain faster.

Freemium also forces a healthy discipline and provides a flow of data for identifying which market segments and which product features have the most value. Once you know who is willing to pay, and what they value most, you know where to put the toll booth. Precision in this is what freemium is all about. If you get it wrong, the flow is reduced and the business suffers and people get angry. If you get it right, the flow continues and people happily get what they pay for.

I hope we can continue to find ways to incorporate a freemium model at Acquia. It's easier said than done, and I'm sure we'll make our share of mistakes along the way, but I really believe in the approach.

Chris (not verified):

Just out of interest what conversion percentage from free to paying users do you guys see with Acquia and Mollom?

Daniel Johnston (not verified):

Freemium is a really interesting strategy. Sadly, I thought it would be a good idea to do something similar in a niche industry. I haven't figured out how to successfully monetise it yet, as I'll need conversion rates of >50% rather than <1% to make the whole thing work. Oops.

Josh (not verified):

Any way the "Acquia community subscription" could stay free?


Chris Ritke (not verified):

Thanks for this well thought out and informative post! All this talk about freemium is really driving me nuts. In a good way. You just made me think about the model for my offerings again, as did Peter's post. Large file transfers and storage are at the core of what I'm working on, so I'm having a hard time figuring out how freemium would apply. With the cost for bandwidth and storage falling fast (I use AWS) and you discussing the merits here, I'm starting to drift back into the 'freemium makes sense' camp again.

JohnForsythe (not verified):

The problem with the Freemium model is there's always someone with a bigger budget that's willing to give away what you charge a premium for. See Yahoo Mail Premium vs Gmail, PlentyOfFish vs, Craigslist vs Ebay, etc, etc. It's generally only a matter of time, and Flickr may well be next.

BryanSD (not verified):

I think one of the difficulties with the freemium model for the client is trying to figure when it is best to make the move from a free customer to a paying customer.

With Mollom, the time to be a paying customer is relatively easy to determine...when your site has more than 100 legitimate posts. With Acquia, there is a larger range of business decisions that are involved in determining when you might want to become a paying customer. Though I think the Acquia FAQ has the right idea in that it might be best to subscribe before you have an issue and not afterwords.

Christoph Weber (not verified):

Very interesting and thought provoking post!
As a web developer using both, Acquia Drupal and Mollom as core offerings to clients, here's my take on the services:

For my own site the free versions are sufficient right now, and the value I get is twofold: The direct benefit of two well designed services, and a chance to kick the tires in a real life situation. As a result I can recommend and use both without hesitation in client projects.

One of my upcoming projects is a real estate agent site. In this business unfettered access for site visitors is crucial. Wide open contact forms, wide open sign-up for services, comments without user registration or moderation, etc. In other words, Mollom is mission critical for protection and therefore high quality leads, and I expect a real estate agent to see the value and pay for the service in exchange. They do this all the time for other services, nothing new here.

So, it is a win-win-win situation: Mollom wins paying customers, I can test, use and offer a critical feature, the client stands to gain more business with less effort to weed out junk and spam. What's not to like?

Christoph Weber

Spenser (not verified):

Having used both completely free and extended free trial, the completely free is the more difficult to convert to a profitable business. The userbase is biased towards believing that the service should remain free forever.

Even when the user is a Fortune 100 company, the individual controlling the account has often opened it as a means of bypassing purchasing bureacracy.

Tim De Coninck (not verified):

Hi Dries,

I think you might say that you succeeded in making it work - doing good and well at the same time. It would be good however if you could give us some insights on how you're currently looking at the different variables and questions since and what kind of valuable lessons you have learned.

All the best,