Design a new web site or a new application

 

I was looking for best practices to create a web site or in my specific case a facebook apps design.

This is for our MyNewResolutions Facebook applications. See this blog post for more information about it.

It seems that the most common way is to use Photoshop and then to create the XHTML/CSS from there. This is most common for graphic designer.

For HTML hackers, it is easier to start directly the mockup in XHTML/CSS.

If you are neither a photoshop guru (or an artist) nor a HTML hackers, you need to use a Fireframe software with a wysiwyg interface. The most common one seems to be from Adobe. Then, you still need to transform the resulting design to proper XHTML/CSS.

I was looking for a super simple and super quick Fireframe application. Facebook apps is usually not high in graphism anyway. Powerpoint is one solution but it seems to me cumbersome for web design.

And I found Balsamiq. I think this is the perfect match to quickly design a Facebook app. It is also possible to create automatically the XHMTL/CSS from the mockup with Napkee or BmmlExporter. I did not try those applications though.

Outsourcing an application in India

 

For our MyNewResolutions Facebook applications (See this blog post for more information about it) we were trying to understand the possibility of outsourcing the coding.

In my quest for Indian programmers, here is a quick study of the best outsourcing web sites.

Outsourcing website with Alexa ranking:

elance.com - 433

oDesk - 639

GetAFreelancer.com - 1200

rentacoder.com - 1800

guru.com - 1900

http://scriptlance.com - 3000
http://www.mturk.com - 7700

TopCoder - 35000

www.globalworkforce.com - enormous in US

And some review from various web sites:

www.elance.com – Generally a bit more expensive than other web sites, but still much cheaper than hiring someone locally in the USA. Providers on Elance are often of a higher quality than other web sites.

www.guru.com – Another possibility with freelancers in many different fields such as engineering, sales, legal – this site doesn’t focus only on programming.

www.odesk.com – Odesk is a bit different from other freelance websites in that it is more geared towards hiring people on an hourly basis and hiring teams of people. This is more effective for long term projects.

www.rentacoder.com – Very good for programming projects done at a fixed price. It works by escrowing the money for the project to give the provider some security, but you don’t need to release it until the project is completed.

www.globalworkforce.com – A provider of full time employees from multiple different countries (the Philippines, India etc).

 

Summary

I used elance and odesk to get quotes but finally as we were in a rush we had to build the application ourselves.

I got 4 propositions for the full facebook application: $250, $500, $500, $460

This was for a simpler version of MyNewResolutions. In our opinion to do the full MyNewResolutions Facebook application, it would require at least 5-10 times this.

15 propositions with hourly rates. From $6 to $20 per hour. Those prices are dependant of the number of referral (with $6, the engineer has no referral). The average is around $9-$12.

The birth of a new Facebook application

 

After many months (year?) of discussions about creating a web2.0 application, we have finally done it.

Here is the baby: Mynewresolutions on Facebook

to be fair, I had a couple of friends super motivated by this idea and I thought it was a good time to jump on the wagon. Here are some tidbits from one month of work.

The idea

The idea behind Mynewresolutions is very simple: it is the end of the year, people need an application to log their new year resolutions.

If this is successful, there are lots of possibilities to keep people interested by actually keeping track of those resolutions.

Design and framework

It is just so easy to jump into the code without thinking too much about the usability or the design. Especially as we were under crunch time and we thought the application would be small.

Big mistake… we wasted a lot of time on changing the design. It is almost as doing the application 2 or 3 times.

See my specific post about some other thought gathered on the net:
http://zhorba.blogspot.com/2009/12/design-new-web-site-or-new-application.html

Coding a facebook application

No beginner guide here. There are so many already on the web. We were almost beginner developers one month ago. The last time we coded was 10 years ago during our study. And it was a long time before social networking. This application is the fruit of 200-250 hours of work with all the learning.

We were under crunch time (the end of the year is only… once per year) so outsourcing was not an option. We did explore the idea though:
http://zhorba.blogspot.com/2009/12/outsourcing-application-in-india.html

The good news is it is so much faster than before to build a good application. Ajax is now quite easy thanks to Jquery. And the APIs and examples are everywhere on the net.

The bad news is about the Facebook API. Very sparse and bad documentation. The API itself is changing all the time and very buggy. This means applications have to be updated regularly and books are almost immediately outdated.

Monetizing

This is also a bad news. It is now very difficult to make any money from a facebook apps.

The success of an application is now related to its “stickiness” or ratio between daily users and monthly users. 

This is thus no surprise that most successful applications are games.

This probably explain the relative mediocrity of non-games applications on facebook. No revenue means no investment in those applications.

Trends for monetization:

1) Advertisement is not working. tons of data on this subject on the net.
2) Affiliation – Seems not that successful neither – No link about revenue from this model.

3) Virtual currency/virtual goods

4) Resell Emails address

5) Resell the application based on the number of active users. Even with no revenue those apps can be used in a marketing campaign to increase the users of another social media.

Almost 4 percent of applications have over 10,000 monthly active users which isn’t bad (June 2009)