Monthly Archive for December, 2007

Another example of Persuasive technology – Frozen Pea Fund

The Frozen Pea Fund has been set up to persuade people to donate the price of a couple of bags of frozen peas every Friday for Pink Ribbon Breast Cancer Awareness.

Interpreting feedback – users and metrics

Daniel Burka is the Creative Director of Digg and the Co-creator of Pownce. Digg’s been going about 3yrs and has a user base of 2million, whilst Pownce has been going about 4 months and has just over 100,000 users.

Neither Digg nor Pownce got where they are without taking onboard feedback, interpreting that feedback and folding it back into iterative releases of the product on a continual basis.

This kind of development process involving constant iterative improvements is exactly what Facebook applications need in order to spread and gain in popularity.

The Standford students that developed the SendHotness application on Facebook created one of the most popular and successful applications in the class. Currently, it has attained over 6 million installs and averages about 125k daily users. Pretty mind blowing for an application that is about 7 weeks old.

Their secret: They spent at least 80% of their time on metrics. Create the viral loop, measure it, test it, change it, tweak it…Repeat until your brain goes numb…Simple assets such as images really made a difference in influencing people and getting more people to download and use the application.

“…it was all about the persuasive nature of our changes”

One interesting insight they mentioned in an interview was the dramatic improvments in sent invitations when they chose to do away with a large invitation window with thumbnail images of freinds and instead have a minimized invitation window with no images; psychologically, the feeling from a user’s perspective when seeing the large invitation window was “oh no, not another invitation screen – I hate sending invitations [skip]…”

Facebook weekly changes

Facebook’s changed the rules again and these ones are worth noting.

3 things have changed since Tuesday…

1. Notifications and Invitations on your profile page have now been bundled together. Application invites now have less real estate on people’s profile pages than they did a week ago. This is bound to be a good thing for people; not so good for application developers.
requests vs. requests new

2. NewsFeed ranking. Feed stories published onto a person’s News Feed which requires that person to download and install the app before reading the content will be penalised. This is analagous to web sites which force you to create an account before viewing content and I’m sure a welcome change – but not for developers who used it to increase their download base. Read more here.

3. Freind lists. You can now group your friends into personal silos for invitations and notifications. Profile page settings based on these groupings are sure to follow so that work colleagues and perspective employers can only see some aspects of your profile whilst close friends can see the full gory details.

Facebook – a persuasive technology

B.J.Fogg is a professor at Stanford University specialising in Captology, the study of persuasive technology. He used his learning in the field of Captology for a course that was run at Stanford for Facebook application developers recently.

Over the course of 8 weeks he and his associates taught how captology could be used to create successful Facebook applications. The premise for creating and teaching the course was based on the hypothesis that Facebook is, at its heart, a persuasive technology utility platform.

Facebook’s foundations are based on your real off-line friends and families. The ones you trust, the ones who recommend products and services to you when you’re at dinner or whilst having a drink after work and the one’s whose views are based on their own experiences and opinions, which you trust.

Collectively the Stanford facebook applications numbers were as follows:
Over 16 million installs in 10 weeks.
~ 925,000 Daily Engaged Users

The Facebook Application Install Stats were as follows:
* 5 apps had 1 Million + installs
* 10 apps had 100K+ installs
* 20 apps had 5K+ installs

The Facebook Engagement Stats were as follows:
* 6 apps had 100K daily users
* 10 apps had 10K daily users
* 20 apps had 500+ daily users

Total Valuation per Adonomics numbers = ~10 Million Dollars

These numbers are incredible considering that each team comprised of 3-5 people…

There’s seems to be alot that can be learnt by applying Captology to Facebook applications.

Differences between MySpace and Facebook

I read over the weekend a piece in The Observer (on page 3) to do with Facebook. It talked about the ‘friendliest’ person on Facebook in the UK. I was slightly taken aback by the term that has been coined to refer to Facebook users who have 1000’s of friends: ‘Facebook Whales’.

People with large numbers of friends on Facebook are seen as lonely, sad individuals. Compare that to MySpace and the perception that friends = popularity…It struck me how different these 2 Social Networks are, made all the more distant by this illustration:


Some thoughts on Facebook apps for brands

recently_pop There’s been some great discussion and insight recently into Facebook applications and how to create successful ones. The Stanford class’s reach of a combined 16+million Facebook application downloads in 10 weeks is testamount to what can be achieved with alot of hard work and determination and a surprisingly small development team.

What has become clear from their success and insights is the emphasis on metrics to substantiate what persuades people to download and use a Facebook application and how you can use metrics to keep track of your ‘persuasiveness’ – It’s interesting how the professors of the Stanford Facebook course describe Facebook as a “Mass Interpersonal Communication” tool.

There is a science to creating a successful Facebook application: Constant testing, measurement, metrics and application iterations are key.

18 Insights for brands engaging in Facebook (and now Bebo) with an application.

1. Set metrics and objectives upfront. Based on the metrics establish why people are engaging or not engaging and tweak your application to improve these rates. Based on the objectives, decide whether your application has been a success.

2. Is your app viral? Generally, the more complex the app is, the less viral it is. The simpliest ones spread the quickest. The more complex the app the more time they spend on the app’s canvas pages.

3. Go in light on the branding: A Facebook application can be “sponsored” by a brand and still attain marketing goals. Don’t beat people over the head with branding.

4. Build applications that “Look and Feel” like facebook. People engage with familiarity of design. Don’t have graphics with colors that overpower or clash with the facebook frame if possible. Don’t use rounded tabs for navigation. 95% of facebook applications use the same style of navigational tabs. Do the same.

5. Plan on A/B split testing & measurement for ongoing tweaks of your application.
The key metrics are Keep Rate, Drop Rate, Install Rate, Uninstall Rate. Others include the amount of time people spend interacting with your application’s Canvas Pages.

6. Metrics are critical. Many application teams set up custom metrics beyond standard Google Analytics in order to measure very specific details of an application. You need A/B split testing on things such as page layout, images, text, user flow etc. You can get enough statistical info within 12hrs to measure the effectiveness of different iterations of your applications.

7. Change 1 thing at a time for iterations and measurement towards desired metrics.

8. Pay very close attention to any trend or change in behavior; Un-install rate increasing for example. This can point you in the direction of a problem, either slow responsiveness (a scaling issue) or something being broken (a development issue).

9. Wording is very critical for influencing behavior. Change and measure. Part of the “Mass InterPersonal Communication” concept.

10. Pay close attention to the experience flow of how you pull people through your application and force desired behaviors to occur.

11. Consider “levels” or “rewards” as an incentive for participation. This means unlocking specific features of the app when a person takes desired behaviors.

12. Invites are critical. Do it right. Most facebook applications do not do it in the most effective manner.
A note on this is that Facebook has bundled invitation (which appear on the right hand of your profile page) together, giving only the most recent invite a prominent place on your profile.

13. Look at how you can easily expose behaviors through the viral carrier mechanisms available (mini-feed, news-feed, notifications, email, participatory sharing, etc.) without the user not having to think about sharing information.

14. Try not to shoe-horn a current app that’s successful on to Facebook. Try to make something totally unique to Facebook. Use the ‘Social Graph’ that is at the heart of Facebook to facilitate new and unique connections between people and their friends on Facebook. Give it a Facebook ‘twist’.

15. 2 attributes that make a successful widget are Attention and Interaction, also known as Engagement and Viral Growth. Pay attention to both.

16. Plan for continual evolution. Facebook’s Platform is in constant change with new updates, features and restrictions added almost on a weekly basis. Developers aren’t notified of the changes by Facebook unitl they occur and you need to be able to react. Application creators need to adapt to those changes and challenges.

Have a plan in place for revisions and additional feature sets. The application will evolve on a continual basis mainly due to people’s feedback. You need to incorporate this feedback and build it into future versions. It shows that you’re listening and creates affinity between your application and the people who use it.

17. Plan for scaling. Budget accordingly. It’s best not to build an app in the expectation that it will be very popular but instead have a plan in place for scaling the application over the short to mid-term. Applications can spread virally incredibly quickly and you need to be able to react to that. Consider using databases that can scale according to use (Amazon’s S3 is a good example)

18. Competitve landscape.
Look at you’re Facebook competition. Keep up to date with the changes they’re implementing and react accordingly.

These points are based primarily on learnings from the Standford Facebook class and on Facebook application development insight from Rodney Rumford and Jeremiah Owyang.

Get Adobe Flash player