Harness the Power of 'Freemium' Learn when--and how--to start charging for your web service without starting a riot.

On the Web, services such as Google, Facebook, and Twitter are the undisputed kings of free, providing robust services at no charge to users. But a host of other companies use free in a much more traditional way--as an enticement for paid upgrades. The model is known as freemium, and while it may not make headlines, proponents say it can make millions.

Unfortunately, finding the right balance between free and paid in the freemium model can leave a new business in uncharted waters, with few hard and fast rules to follow. Here, three entrepreneurs using the freemium model with services in different stages of development explain how it's being implemented in their businesses, and how it might be implemented in yours as well.

Set expectations from the beginning
If paid web services have an evangelist, it may very well be Jason Fried, partner at 37signals, a web-based software company that offers a suite of online applications. Basecamp, the company's flagship product, helps companies of all sizes collaborate on projects with their clients through a simple but powerful web interface. The privately held company is tight-lipped about revenue, but with more than 3 million user accounts in Basecamp alone, Fried says that both revenue and profit are in "the multi-million dollar area."

As for when a web service should start charging, Fried's answer is simple: immediately. Charging from the outset tells users that the product has a specific value, Fried says. Trying to charge for something that was previously free can undermine the product's value, causing potential customers to ask why the service is suddenly worth more than it was in the past.

"My feeling is that you should begin charging right from the start...The longer something is free, the less it's worth," he says.

Chris Nagele, principal at Wildbit, offers a slightly different perspective. Starting as a free service can work, he says, if the addition of paid plans is done correctly. His company's flagship product, Beanstalk a version control service that helps software developers and designers track and save changes to a project, launched with only a free plan, but that free plan came with the understanding that paid plans would soon be available.

When the paid plans launched, the free plan was left unchanged. The new plans simply offered more features for a higher monthly fee. Adding a price only works if you're adding value as well, Nagele says.

Scratch your own itch to find a compelling product
When creating a service that users will pay for, the best source of inspiration is often to solve your own problem, according to Fried.

37signals began as a web design firm, but in working with clients Fried and his partners became frustrated with existing project management software. They developed Basecamp for their own use before offering it to the public in 2004. As a result, the company stopped doing contract work a year later. It now devotes 100 percent of its efforts to developing and supporting its own applications.

"We consider ourselves the target, and then we go out and find other people who are exactly like us," he says.

Sachin Agarwal, CEO of Posterous, tells a similar story. He and his co-founder, Garry Tan, wanted an easy to use, premium blogging platform for their own use. The result was a service in which creating (or updating) a blog is as simple as sending an email. (Try it by e-mailing post@posterous.com.) Attachments of almost any kind--images, videos, audio--are automatically converted and posted along with any text.

The 15-month-old startup has yet to generate revenue, but plans to introduce paid features to complement its free offerings, Agarwal says.

Find the natural break points
Unfortunately, creating the product is often the easy part, at least for those companies that already possess adequate technical talent. Splitting a product into free and paid tiers that provide sufficient value to users while turning a profit--creating a sustainable business, in other words--is more difficult, and sometimes comes down to trial and error.

At 37signals, Fried says, he and his partners look for the natural "break points" in a product, and create tiers accordingly. Basecamp, for instance, specializes in project management, so each paid plan is distinguished by the number of projects it allows a user to create. The free plan limits users to one project. The Basic plan, at $24 per month, increases that limit to 15 projects, and so on.

"The best advice I can give regarding pricing is this: Have a price. Don't be afraid to charge for your work. And make it a number you'd pay yourself," he says.

With Beanstalk, plans are differentiated by the number of storage repositories that users can create, and the amount of overall storage space available to them. The free plan limits users to one repository with 100 megabytes of total storage, while the first paid plan ups that to 10 repositories with 3 gigabytes of total storage for $15 per month.

Beanstalk's paid plans are further separated from the free plan by the ability to deploy completed software to production servers, and the use of encryption for greater security. The latter two features have been the most common reasons users cite when upgrading, Nagele says.

The company did quite a bit of research into prices at other services and did its best to estimate potential costs when creating its price structure, but in the end, some of those decisions were simply educated guesses, Nagele says.

"There's no real how to for it, because there's so many conditions," he says.

Posterous plans to charge only for those advanced features needed by businesses and the most technically advanced users, and for storage space greater than that offered in the free plan, according to Agarwal.

Understand who will pay
Nagele says that some decisions about pricing are directed by the audience that a particular service is using. Services that target businesses or professional audiences usually have an easier time converting users to paid plans than services that target consumers.

Beanstalk currently converts "upwards of 10 percent" of its users to its paid plans, Nagele says, no small feat considering that rates as low as one percent are not uncommon. The service targets small to medium sized businesses, and it became profitable soon after its launch. Its most expensive plan costs $200 per month.

Fried declined to give exact numbers on the percentage of paid accounts among 37signals products. But he did say that "free accounts outnumber paid accounts," and the company has "more paying customers who started as paying customers than paying customers who upgraded from free plans."

The goal at Posterous has always been to serve as large an audience as possible with the free product, Sachin says, while charging a minority of users for advanced features. Though some question the value of serving nonpaying users, he sees the larger audience as a way to inspire development of new features.

"That's where we drive a lot of passion for the product from...That's what makes us happy, what we would call success."

When considering which features should be free and which should be paid, Agarwal likes to use "the mom factor."

"Does my mom need this for Posterous to be useful for her? If so, I think it should be free. If a feature is only important to a smaller group of advanced people, then it's something we can charge for without making the service less useful for the masses," he says.

Focus on the basics
As a startup that plans to serve large numbers of free users, Posterous keeps costs as low as possible, Agarwal says. When it was time to lease office space, the company passed on the trendy area next to Twitter's offices in San Francisco, settling instead on a much less expensive space in the North Beach area. Posterous is operated by three full-time employees and one contractor, and the company intends to grow no faster than it needs to.

Fried says too many businesses--both online and off--get distracted with things that don't matter, whether focusing on wasted employee policies, scheduling unnecessary meetings, or worrying too much about branding. What's important is to build a great product that's worth more than you charge for it, he says.

For Nagele, it all comes down to value.

"A lot of this stuff is really just fundamental old school business models, you know. We're entering a space where a lot this stuff seems unique, but in the end it's all about how much value you're offering, and if you can offer enough value people are probably willing to pay for it," Nagele says.