Tuesday, December 20, 2005

Adaptive Path and online cruise booking

Today's edition of the Adaptive Path email newsletter contains this highly interesting "Case Study: Princess Cruises, to Build or Not to Build?"

I don't think they'll mind my re-posting the entire article here, increasing its exposure to the public.

This report is enlightening, but also is a promotion for the services of Adaptive Path, from the information architecture experts and blog pioneers, Peter Merholz and Jeffrey Veen.


Princess Cruises came to Adaptive Path in 2004 with a research question:

Should Princess build an online booking system for cruises?

We helped them investigate and then ultimately reframe the question to encompass the larger goal of supporting individuals throughout the cruise researching and purchasing process.

Through our work, we discovered that booking a cruise online is much more similar to buying enterprise software than booking an airline ticket.


An overwhelming majority of those who book cruises do so through a travel agent, but Princess wondered if there was a segment of the population who would be more comfortable using an online engine.

Defining the Research Project

Princess needed to understand the entire decision-making, purchase, and travel preparation process that the average "cruiser" undertakes, starting from when they decide to take a vacation through to how they connect with new friends after returning home.

We worked with Princess to define and conduct their research into understanding and documenting the cruising process.

Research Broadens the Question

From the research interviews, we analyzed the tasks that a potential cruiser sought to accomplish, and then assembled and documented a detailed mental model of the cruise-buying process.

We discovered through our research that there were actually over 30 variables that accounted for why someone chose a particular cruise and which cruise they ultimately picked. Though not every cruiser used every variable, some significant combination of these affected each decision. Variables ranged from expected concerns, like cost, timing, and port of departure, to more nuanced considerations, like cost of airfare to disembarkation point, and cruise ship amenities.

Once we understood all this, we worked with the Princess team to answer its question of whether to build an online cruise-booking engine. In the broader context of our research, we determined that the core problem was actually that existing cruise-booking engines like the ones used by Orbitz and Travelocity used an airline ticket booking metaphor, which was inappropriate for the cruise buying process.

When buying an airline ticket, there are a smaller set of variables that influence decision making, and the decision can be made in as little as five or ten minutes -- that never happens with cruises. Moreover, many people who purchase airline tickets do it for themselves or a relatively small group (like their family). A smaller variable set and fewer people to confer with significantly simplifies the process.

Discarding the Conventional Model

Given this new development, we searched for other types of purchase processes that had more parallels to the cruise purchasing process -- ones that considered the timeframe and number of people involved, and also provided good information and support before and after the point of purchase.

Princess was trying to decide whether to build an online cruise booking engine, like the ones that Orbitz and Travelocity had built, based on the airline-ticket purchasing process; however, the large travel engines they spoke with were measuring success based on whether someone completed the purchasing process, thus generating dismal success rates. Our research indicated that Orbitz and Travelocity were measuring success incorrectly, and that a more effective process would support people through the decision-making process, and recognize that the act of committing money to a particular cruise is actually the mid-point, not the end, of a customer's online relationship.

Recalling similar situations we worked on with past clients, we concluded that purchasing a cruise was much more similar to the decision-support model used in purchasing enterprise software. Though the products they support are much different, the purchase points and considerations have significant similarities.

If Princess built a typical online booking engine, they wouldn't be any more successful than the online travel sites that had already done that. Instead, they needed to redesign their website to support each step of the cruisers' decision-making process.

Content Analysis and Visual Prototyping

Based on our mental model of the cruise buying process, Adaptive Path performed a content gap analysis to determine what content Princess had, what they needed to produce, and what they didn't necessarily need online in order for both new and experienced cruisers to accomplish their goals.

We also recommended a navigation scheme based on a weighted approach that took into account many of the facets of the decision-making process. We framed the high-level navigation around the key decision points in the researching and buying process, including support after the initial purchase, when travelers want to purchase cruise excursions or select a cabin, and a post-travel section where previous cruisers could access photos and other post-cruise information, and even select their next cruise.


After developing a visual prototype for a redesigned website, we presented the design ideas in conjunction with the research findings. Princess's executives appreciated that instead of simply answering the initial question, we took it a step further, reframing the question to get at what they really needed to know.

In late 2005, Princess launched their redesigned website based on the work we did with them. The site has already vastly exceeded Princess's expectations for both customer satisfaction and revenue. You can see the new Princess web site at:



My first impression of the Princess site is good, except I'm not a big fan of sequential home page images, especially as the head of the home page. I keep wondering if other things on the site are changing as the image changes. It's distracting enough to stall my progress through the site.

[signed] Steven Streight aka Vaspers the Grate

No comments: