Thursday, 12 July 2007

Measuring and ranking Rich Internet Pages

Does the current model of measuring page popularity, based on the number of page hits fit the single page interfaces RIA applications have? Answer: no.

Nielsen/Netratings (Global leader in internet media and market research) are currently going to change their way of measuring page popularity.

For more information on this, use the following links:

Marketing Pelgrim: The problem with measuring time spent on a web site

The New York Times: Nielsen Revises Its Gauge of Web Page Rankings

Friday, 6 July 2007

Rich Internet Applications in your company: Things to consider

Why should you use Rich Internet Applications in your company?

The answer is quite simple: because it's better. It solves a lot of usability issues and technical shortcomings, especially on your webpages. It is even so much better that applications can be created for a web browser which earlier would be developed as a desktop application because of the limitations of the web.

I am not going to tell you what the advantages of RIA's, you can read all that everywhere, for example on Wikipedia.

Here are the things you should consider when making the step:

How service oriented is your company?

This is a technical thing. Maybe you've heard of SOA (Service Oriented Architecture) or SOE (Service Oriented Enterprise). It basically means that you will take the front-end (the interface) apart from the calculations, rules and processes (business logic). These are going to be seperate applications: services.

This means that any application can use a specific service to get it's data. For example a service to give pricing information on you products. An application can use this service to get product information without doing any database stuff or permission stuff or anything. It doesn't have to look if the product is in stock or anything, the service does that.

The services can be introduced step by step and old applications can be rebuild to use these services and so remove their own business logic. So your legacy can stay or fade away in time.

How can we develop RIA's?

Currently a growing number of developers are learning how to create RIA's. The leading technologies are Adobe Flex (based on Flash) and Microsoft Silverlight. Research of The Butler Group in June 2007 state that also Nexaweb and Sun Microsystems are major players on the 2008-2009 market. You should also consider Backbase, edge IPK, OutSystems and TIBCO.

Your developers will be more front-end or back-end oriented and they should speak eachothers languages to be able to communicate.

The way a RIA project should be run differs from a standard ICT project. It should be user oriented, business oriented and technology oriented. In "classic" ICT projects the user usually is forgotten. To improve the user's experience this is neccessary to make the project to a success.

This takes care of problems you usually experience after the product is launched. Redesigns, remakes and fixes then usually cost 80% more and doing a user centered design takes care of these things in an earlier stage, when changes are much cheaper.

Will my users understand this new stuff?

Yes! Ofcourse currently there are not a lot of RIA's online, but the number is growing. Because of the user centered design, the learning curve for the user is short and the benefits are instantanious. Maybe your users won't notice that they are working with a RIA, but they will be happy about how everything works.

Don't do too much!

Don't make your website one big RIA. You will run into big problems (Google-ability, print issues, etc.). Take the "application like" things, like the order and checkout process, searching and selecting things or processes with a lot of forms or form-elements in it. Make a hybrid website (part HTML and part RIA applications).

Okay, nice. But where should I start?

Start with your most important website or the site a lot of people complain about. Make an inventory of the problems. Let a Usability Analist (like me ;-)) do a Usability test and let him/her make recommendations. Then you can start considering technology things to work on those recommendations.

And do not forget: Define what your goals are! State how much more products you want to sell, by what percentage you want to decrease the number of helpdesk calls, etc. Quantify them!

Do testing in early stages. The sooner you discover problems the better (and the cheaper to change). Also test afterwards and monitor if your goals are met and why or why not.

Don't start with too large projects. I can tell you a hundred things about how to do things, but each company has it's own politics and ways of doing things. You can get accustomed with that in the first smaller projects.