I’m looking forward to participating in the Garage Technology Ventures’ “Revenue Bootcamp” on July 10th, discussing, amongst other things, freemium business models, digital goods, and affiliate programs.

What is Freemium?
A business model where a company combines both free and paid services. The free can be:

  1. A feature-limited version of the site, with a feature-rich version available for upgrade
  2. A “try before you buy” version of a product, with a required time-period to buy the product
  3. A customer-type limited version of a product, where certain customers qualify for a free version, and others must pay
  4. A seat-limited version of a product, where a certain number get the product for free and over that number, payment is required

Examples in the real world:

  1. Feature limited, versus premium: Photobucket, Flickr, LinkedIn, YouSendIt and many, many others.
  2. Trial version: Quicken.com, Audible.com, Rhapsody.com, Pandora.com
  3. Customer-type limited: BizSpark from Microsoft
  4. Seat-limited: SocialText free for up to 50 users

I’m going to focus here on the first type of freemium. That’s the one I’m most familiar with, and the model adopted at Photobucket.

In my experience at Photobucket, having a free offering, and a super-simple, quick registration process, allowed the company to welcome tens of thousands of new users every day, day-after-day. This large registered user powered a terrific network effect. The free offering at Photobucket is subsidized by advertising and print services.

When does freemium work?
A freemium business model can work when the following conditions are true:

  • You have a free service that offers enough real value that you build a large, active user base of free users that you can market to
  • You have a good balance between free and paid features. (As Josh Kopelman of First Round Capital says “Too many freemium models have too much free and not enough mium.” So watch out for this!
  • The price point you choose for paid features match the willingness your target premium user segment are willing to pay (see “What to charge for premium” below).
  • You practice good marketing techniques at every decision point to convert free users to paid users — it’s like a friendly “bait and switch.” Hook them for free but work hard to ‘sell up’ to paid.

Some advantages of a freemium model:

  • There’s a reduced risk for the buy – it’s free! Users will try it because it’s easy to commit.
  • Free has shown to be a great way to get lots of customers in the door! Photobucket has over 80 million registered users for their freemium service – that’s a big base to market to!
  • The free service becomes a fantastic test-bed for new features and building a better product – early on, users have a high-tolerance for bugs if they know it’s free. However, conversely, when a free product becomes mature (certainly in the case of Photobucket) high standards are expected by users. At Photobucket, we used a totally different site – www.tinypic.com — as a feature test-bed site. Photobucket had become so mature that our users didn’t want to be experimented upon too much. So we tried things out on Tinypic, and, if they worked there, moved the code to Photobucket.

Evaluating freemium: Revenue
It’s mandatory to clearly calculated advertising revenue to your free base, to your other sources of revenue. Some calculations to make:

  • Active user rate, or the percentage of your registered user base active in a month.
  • ARPU (Average Revenue Per User). To get your average monthly net revenue per free and paid subscriber, divide your net revenue in that period by the average number of subscribers in that period.
  • Churn, or how many users you are losing in a period. To get your churn rate, divide the number of subscription terminations by the total subscribers at the beginning of the period. For example, if you get 10,000 subscribers on January 1st, with 100 terminating in that month, your churn is 1%.
  • Your advertising revenue/user, compared to premium. For example: Your site has $1.00 CPM. That’s $0.001 revenue for each user’s visit. Say a user visits an average of 4 times/month (or 48 times/year). That’s $0.048 in revenue/user/year from advertising.

Once you have these numbers, look at your active user rate as a percentage of your overall users Then estimate upgrading 1-2% of your active users to premium as a start, and model the result.

Evaluating Freemium: cost
Do you have a break-even, or close to break-even model for your free users? In other words, is your advertising revenue pretty much offsetting the costs of offering your free service to your users? You have to know this. It’s healthy when you are pretty much keeping your marginal costs close to zero (costs are offset by advertising) because then you can really focus on your premium services as your money-maker.

If your marginal costs are significantly above zero (it costs more to offer the free service than is offset by advertising), then you have to look at your free service as one big marketing funnel to drive people into the premium service. In this scenario, you may look at free service more as a cost of acquisition, and a cheaper cost than other methods. This can also be healthy, but it’s critical that you focus major efforts on pushing people up to premium.

Expected conversion rates
Historically, people have considered getting 1% of your installed base to upgrade has been healthy. But many companies are seeing much higher percentages of upgrade. Longtail.com reports these interesting numbers:

  • Club Penguin: 25% monthly uniques pay, $5/mo per paying user
  • Habbo: 10% monthly players pay, $10.30/mo per paying user
  • Runescape: 16.6% monthly uniques pay, $5/mo per paying user
  • Puzzle Pirates: 22% monthly players pay, $7.95/mo per paying user
  • Flickr: 5-10% monthly users pay, $29.95/year
  • Ning: 3% of its 500,000 social network creators pay for premium
  • Xing: 8% of its 748M user base upgrades to premium services (see Xing profile below)
  • Shareware: typically less than 0.5% of users pay

Encouraging upgrade to premium

  • Don’t make the “free” version too rich
  • Survey your most active users, and find out the features they find most valuable on the site
  • Get in front of users and talk to them! Any way you can. In person. Via Twitter. On profiles. Ask them what they’d pay for.
  • Be up front, and explain why some things are charged for. Most users understand that free can’t include everything. You’ll lose some users who will never, ever pay. But you’ll endear yourselves to those users who are willing to cough up a buck or two if they understand why.
  • At every decision point — even at first registration, educate on the fact that there’s an upgraded, premium service and what’s in that service
  • Testing, adjust, test again to find out the messaging that is working — do your users want to be identified as a premium user to the world, do they just want the feature, or both? Keep iterating.
  • In lists of features and functions, try ‘greying out’ the features that are unavailable to free users — they see the feature, but they can’t ‘get at them’ without upgrading
  • Make sure users know when they’ve reached their limits, or are close to reaching limits; at Photobucket, if bandwidth is exceeded, then all images posted on other sites are replaced by a “bandwidth exceeded; upgrade now” message. Harsh, but effective.

What to charge for premium?
Here are some simple ways to start the decision-making process:

  • Review your competitors, as well as products in adjacent spaces; compare pricing models
  • Consider opportunities for more than one tier of offering; can you do three-level price point? That may be complex enough for a start: free, medium, and high-end. Give yourself the flexibility to move features around as you measure your user behavior and feedback
  • If you already have free users, conduct a survey of those users. Pricing surveys can set you back tens of thousands of dollars — hard to swallow for a small startup — or even a large one. Take a look at some of the templates available through Mineful, or SurveyMonkey and try them out.

Bloodhound tracking … sniff out every metric that makes sense
The article here “The metrics we track and report: example for freemium products” by BuzzGain is a good place to look at the kind of metrics you should consider tracking. This should become a monthly published dashboard of its own. (See post in this blog “What’s on your dashboard?”)

Profile: YouSendIt
YouSendIt has been successfully marketing with a Freemium model since 2006, and claims 100,000 paid subscriber. In this interesting case study of their freemium model provided by Andrew Chen, they provide the following advice:

  • Adopt the freemium model wholesale and focus on it success. Don’t just consider your model a free service with an ‘optional’ premium offering. Devote your energies to pushing users up the funnel.
  • Users may call you a ‘sell out’ if they’re never prepared to pay. Users who are willing to pay will likely encourage similar users — a healthy thing.
  • Measure everything you can! See metrics section in this post.
  • Work on your pricing model. See what to charge section in this pot.
  • Look for lots of 1% conversion lifts. They’re easier to find than 10% conversion lifts!

Profile: Xing
Xing is a European competitor to LinkedIn, and a public company. So we can look at their reported results. In Q1 2009, they reported revenues of 10.8M EUROS, with their number of Premium users counted as 8% of their user base of 7.48M users (600K premium users).
They have the following revenue streams: upgraded premium accounts, Xing job services, and advertising. While their advertising revenues suffered the same decline almost everyone is feeling, other revenues grew for the company. Job seeking is big business in a down economy!

Profile: LinkedIn
As they are not (yet!) publicly traded, it’s harder to get validated figures about LinkedIn, though a number of articles put their revenue at a profitable $100M, with revenue approximately equally divided between their three revenue streams: upgraded premium accounts, hiring services, and advertising.

You can see how they compare features for free on their site. They clearly focus on making sure connecting is unlimited (thus growing the user base), but if you want to really lean on the email and messaging features, you have to pay for it.

Profile: Socialtext
SocialText, now free for up to 50 users. Press release: (June 23 2009). Their new free offering aimed at mainstream user for up to 50 people in an organization to collaborate using Socialtext’s social software platform.