Website (Re)Design & Kaizen

Time for some more shop-talk. I’m no expert on usability, but something tells me that there might just be something amiss with Ryanair’s website. Unlike most, I don’t have too many quarrels with the airline, but their website is just plain bad – which is especially bizarre given that it’s so core to their business (tickets are only available on the web, or by phone for a premium).

I count an impressive 6 navigation bars, each with it’s own unique styling. Cross-selling is clearly key to their business model and helps make up for the shortfall in providing artificially low priced tickets. However, I don’t see how this overload of offers, poor usability and shocking design can really help them achieve this goal.

Easyjet’s website, on the other hand, provides a much needed respite. This is curious since they share the same cross-selling model, yet execute upon this with completely different strategies. I’d love to compare their website analytics data to see differences in conversions and bounce-rates.

Another example of an e-commerce site with bizarre usability is TheTrainLine; a UK train ticket vendor operated by the Virgin group. It’s usability has been documented by several people (here and there). What’s even more depressing is that the booking engine is employed by other companies through a white-label solution; further spreading frustration and poor brand experiences.

Strangely, Reading Room (a creative agency) scooped an award for constructing the TheTrainLine site. However, the commission has since been scrubbed from their Portfolio page. Archive.org helps us dig up the original page – they can’t escape their past that easily. It’s not all doom and gloom though, as TheTrainLine have seen the light and made some much needed improvements as of late.

While the demand for website design is still high, the supply of vendors is gradually becoming saturated. However, I believe there to be large potential in positioning yourself as a conversion-focused website redesigner. Your mission is not to deliver a new design and do a runner, but instead provide a service of making gradual improvements (Kaizen) to the existing site; with the chief aim to increase conversions and advertiser ROI.

To achieve that end, the outfit will need to avail of split testing methodologies to evaluate the performance of particular changes. Here is a rudimentary overview of how this may operate:

The extension of this is to deploy multivariate testing; whereby multiple changes can be tested simultaneously and an algo steps in to figure out the optimal combination.

The beauty of this business model is that instead of delivering a glitzy new site, you can can instead hand over an optimised site which you can prove (by testing the final version against the original) to drive the business an incremental $X. From this, the firm can comfortably demand a X% royalty. Win-win.

MarketingExperiments.com has an archive of case studies documenting how tiny changes can have a big impact on user retention, conversion rates and ROI. Based on the knowledge they’ve amassed, they also showcase mock-ups of hypothetical improvements for branded sites. When I get a chance, I may have a crack at doing the same for Ryanair.

Time to name and shame. Have any other examples of confusing, shocking, or just plain ugly usability or design?

 Tags: Business, Internet   Published: 1st October '08


6 Archived Wordpress Comments

john

That is bloody ordinary isn’t it.

Keith Mander

What is?

Keith Mander’s Blog | Introducing FixMyFloors.com

[…] traffic as possible that can provide a demonstrable ROI; and improving conversion rates through split-testing techniques I’ve previously talked about. Permalink •  […]

Inder P Singh

Just wanted to point out that multivariate testing may be more appropriate during the early period when the design is being developed. Making groups of changes at one time could lead to faster design decisions.

When the best (or a good) design has been chosen for the website, split testing may be employed to optimize the design, one element at a time. In split testing, a statistically significant (in other words, large) number of user decisions should be considered before choosing one design element over the other.

Inder P Singh

Keith Mander

Sound advice Inder. Thanks for stopping by.

Keith Mander’s Blog | Optimise My Site & Win a Prize

[…] to make the most out of the inbound traffic. I now intend to focus my efforts on the latter. As previously discussed, it’s incredibly easy to develop and roll out simple A/B tests and steadily increase […]