Notes from FacebookCamp Toronto 2, Part 3

The crowd at FacebookCamp Toronto 2.
Photo taken by Joseph Thornley — click it to see the original on its Flickr page.

And now, the last of my notes from yesterday’s FacebookCamp Toronto 2. If you missed the first two parts, follow the links below:

• Notes from FaceBookCamp Toronto 2, Part 1
• Notes from FaceBookCamp Toronto 2, Part 2

Monetizing your Facebook Application (Greg Thomson)

• What does he consider a “failed app”? “An app that has less than 10K active users”.
• What does he consider a success? An app with about 250,000 installs. With that, he says he “could go full time”.
• What’s an active user worth per year? “About $3”.
• “You can’t just throw one banner up and hope to make it.”

Case Study: “My Garden” App:

• Each flower was worth a certain number of coins
• Each user has an allotment of coins — extra coins had to be paid with cash
• The scheme to make money through coin sales didn’t work — “In my experience, users will generally not pay for things.”

• Across his apps — the big one being “My Aquarium” — the installed base progression was from 0 to 250,000 to 8 million installs across apps
• Did not add ads until his installed base hit the 250,000 mark, and he regrets that
• Why wait so long before adding advertising? He was worried that they might drive users away.
• This was not the case: users are used to ads
• There is a hidden upside to starting late: he didn’t get discouraged at the beginning when ad revenues would have been small.
• The money he’s making covers costs of servers “definitely starting to mount”
• “You really need large numbers to effectively monetize applications.”
• Some demographics:
  • 60% of the users generate of 90% of revenue
  • 30% generate zero revenue
  • The country breakdown of his users:
    • U.K.: 20%
    • U.S.: 20% — These users represent the best monetization opportunities — try to increase them!
    • Canada: 20%
    • Other countries: 40%

Revenue basics:

• Appsaholic (Social media) — pay per click — average $0.10/click
• Adsense (Google) — pay per click — average $0.07/click
• Adsense worked well with “My Garden”, since its theme tended to result in relevant ads — ads for florists and the like.
• Adsense didn’t work well for “My Aquarium”, since “people don’t really send each other fish” in real life, and the ads weren’t relevant.
• Incentivize your users to complete offers: give them rewards in exchange for their completing surveys, which generate revenue

Create a new want:

• Most of the popular applications are entertainment-based
• Revnues for apps go through a cycle: start / peak / decline / plateau
• One way to avoid or at least forestall plateaus is to cycle through ads — change them up often
• You have to keep putting new stuff up — both content and ads
• Change things! Vary the order of ads, or run slightly different ads — any variation helps!

What else?

• Cross-promote other apps in your app to generate new users — he does this
• In his experience — he can deliver 50,000 installs for other apps because of his user base
• If your app is a good fit, you have the opportunity to do custom advertising with large brands — such as the “Zombies” app, which got sponsored by Resident Evil
• Use A/B testing to maximize revenue — show 2 different variations of an ad — see which performs the best

Selling your app

• No major acquisitions just yet
• There was talk of the “Where I’ve Been” app getting bought, but it hasn’t happened yet
• Expect to get $1 per installed user or $10 per daily active user
• You should:
  • Have a base of banners
  • Build incentivizable offers
  • Get surveys through affiliate networks
• Note that a lot of offers are not incentivizable

• Up until a week and a half ago — all users on 1 app server and 1 DB server
• There are 10 rendering servers for Flash and jpegs

• What his time to market? “Keep it to a week. More than a week? The odds of failure go up dramatically.”
• How long to 250,000 users? “2 weeks”.
• How long to 8 million users? “Been doing this for 3 months now.”
• His operating costs? For servers — $2,500 a month

Secrets of PayPal interface used by Gift Cards Facebook Application (Steve Pritchard)

Background

• His “Gift Cards” application was a convergence of business opportunities:
  • He had a Toronto business associate with abundance of gift cards to give
  • Toronto has a high density of active Facebook users (remember, it’s the #2 Facebook city after London)
  • Obvious application fit for his business
• The application cried out for PayPal interface
• Wanted to offer a simple payment scheme
• Wanted to avoid complex HTTPS interactions
• “Challenged by the challenge”

Challenges

• PayPal data does not fit on Facebook Canvas
• Required pop-up window
• Had to be done with IFRAME; and the canvas and IFRAME cannot communicate
• PayPal pop-up window must have specified dimensions
• Have to update canvas under all PayPal termination conditions
• Had to sync 3 threads on 4 subsystems

Solutions

• Solution: Get user to click twice:
  • First click to start the PayPal/Facebook sync loop
  • Second click to open the PayPal pop-up window
• Gift Cards server mediates
• Again, a click-twice user interface:
  • First click to start Facebook canvas polling cycle
  • The canvas’ PayPal “checkout” button is replaced with an IFRAME version
  • Then, the intermediate step of asking user to confirm amount
  • Generates 2nd click

Q & A

• How long did it to develop? “I was learning Facebook, I was learning Paypal…so about a week.”