Monday, October 27, 2008

Defining RIA: Borders, a case study?

Rich Internet Application (RIA). What is it? How would you explain it to someone? It is not easy to do since the "rich" part is really subjective. For years I have thought about this definition as part of the Web 2.0 Patterns book and the troubles we will have trying to quantify something as intangible as "rich" (rich is something a lot of us used to be until we looked at the stock market lately).

Today I got a break on it. HB Mok (a colleague at Adobe), sent out a notice that there is a case study done on the Borders.com web application showing they had a 62% increase in conversion! While the Adobe talking points on this were around how Flex was used, I also found this valuable for anyone trying to reinvent their business. Naturally, if this is your interest, you will want to download the Allurent Case Study from here.

I don't want to spoil the entire case study but I will relay some of the more important findings:

o 62% higher conversion

o 41% more products viewed

o 11% more likely to recommend

Borders.com is worth taking a look at if you haven't already.  Is it possible that "rich" from RIA really means that you have the information at your fingertips you need?  Could it mean that your forms load quickly and embrace undeniable logic (example - not asking you for a "zip" code if you're shipping to Canada)?  Could it mean the perfect balance of graphics, data and interaction, all blended perfectly in an intuitive package that speaks to the users?

I have to ask these questions pragmatically.  Flex gives you the power as a developer but you still need to do the architecture behind it.  I applaud Borders.com for their work of art combining logic, emotion and function.  

To the developers and architects at Borders - well done!  

6 comments:

  1. While I don't doubt this RIA improves conversion I think the pseudo-science underlying the "Magic Shelf" is almost as magical as the shelf.

    To be clear, this was not a statistically valid multivariate split test, where customers got one case or the other with their results measured. Rather, customers using the Magic Shelf bought more and recommended more.

    Simply put, this study did nothing to avoid self-selection, where the most active shoppers were also the most likely to use the Magic Shelf. So until a study shows a valid A/B test, I'll assume this is "Magic Marketing" rather than a "Magic Shelf."

    ReplyDelete
  2. Agree. I would not even consider it a scientific study, just a clue of interest. It was also more of an epiphany for myself.

    I wonder how to quantify such a study? I think it would be interesting to see the results to say the least (plus get me over writers block).

    D.

    ReplyDelete
  3. I would suspect that a more scientific study would be create an RIA version, and a non RIA version, and randomly load balance people to one or the other, with the same data going to both.

    Interesting for sure though.

    ReplyDelete
  4. John:

    Your comment is totally in line with what I'd like to do. This gets to the real crux of the problem. How can you make an RIA and non-RIA version when RIA is not fully defined? Chicken and eggs for sure. I fully agree with what you're saying but I have no idea how one could create such a test.

    What could be the differing denominators? Here are just some initial thoughts. Would love others to add to them and maybe we can stage a test.

    1. Dynamically loaded choices for forms based on knowledge of user vs. static choices? Example - one form asks me what country I am from first, then if I select Canada, it offers me choices of ten provinces and asks me for a postal code vs. a state and zip code.

    2. Information that is not relevant is not presented vs all info is presented. Example - if I live in Canada, products and services that do not ship to Canada are not included on the initial view of as choices for me to select vs. everything is made available?

    3. All users must create a profile, albeit simple so the site can dynamically configure itself to the user preferences based on the UID vs a site that is static.

    4. languages and cultural choices and others preferences based on user knowledge are pre-loaded and made visible vs. static.

    5. Redundant steps are eliminated wherever possible. For example, a user logging in and ordering something should not be asked for their name if it is already known. The form should be pre-filled out or better yet, no form presented other than a "click here to change your details button" vs. forms, forms and more forms.

    6. The RIA site should automatically detect and render in an optimized mode for the client platform, vs. a one size fits all method.

    All - Please add more...

    duane

    ReplyDelete
  5. - (AJAX) dynamic page changes vs. static

    - based on SOA (RIA is Service Oriented Client) vs. tightly coupled or standalone client. RIA can access services ad=hoc when needed.

    - visually intuitive

    ReplyDelete
  6. Интересно, а сколько дас пр этот блог. Мне надо проверить. Сайт то интересный этот, буду изучать, я на ваш блог пришел через сайт Adobe. Понравилось.

    ReplyDelete

Do not spam this blog! Google and Yahoo DO NOT follow comment links for SEO. If you post an unrelated link advertising a company or service, you will be reported immediately for spam and your link deleted within 30 minutes. If you want to sponsor a post, please let us know by reaching out to duane dot nickull at gmail dot com.