Defining the Problem: Q&A with Jen Kozenski & Larry Cornett

by May 14, 2006

For part four of our conversation about Defining the Problem through design, I spoke with Jen Kozenski, Senior UI Designer at Google and Larry Cornett, Design Director at eBay about their efforts pushing a complete reorganization of the eBay marketplace.

For context, eBay is the 29th largest economy in the world with 193 million users so changing the information architecture of the site is no small feat. How did a few designers make it happen? Read on…

Q: As you know, I’ve been running a series about how design can be used to define problems, not just solutions. The work you guys did making the case to redesign eBay's information architecture is a model example. How did that effort come about? What prompted you to try and kick it off?

Jen: Over the years the eBay site was growing very rapidly but organically. It became noticeable that sections of the site were being designed by different teams, which resulted in many inconsistencies and no clear connection between areas of the site. Through usability studies, field studies, and log analysis we saw that users were unaware of different sections of the site and could not find information or features they needed when they needed them.

In order to truly understand the extent of the problem we hired a consulting company who conducted card sorting activities and a series of stakeholder interviews that helped clarify and quantify things enough for the rest of the company to understand the issue and ultimately buy into the need for a solution.

Larry: A key survey found that one of the biggest issues users had with our site was that it was so hard to find anything and that it was cluttered and confusing. Another big issue was that it was hard to explore our site. Primary areas -as indicated by the header navigation- were not accessible unless you were already registered and signed in. Which was pretty bad. For example, a guest could not explore selling on eBay until he first registered as a buyer and then went on to register as a seller and then clicked "Sell" in the header. Ouch.

A guest couldn't even explore "My eBay" until registered. A key part of getting others to understand the problem was to frame things in a way that mattered to them. Talking about the confusion and usability issues really was not enough. We had to quantify the situation as a business problem -that we were losing potential new sellers and that we were losing potential bids.

Q: So following this data gathering stage –card sorts, usability, etc.- how did you pull things together to illustrate what was going on? Was there a specific format you used? Were there any artifacts that really drove the point home?

Jen: In order to communicate the problem to decision makers we had three primary deliverables. First, we presented diagrams of the users' mental models –which were based on the result of the card sorts- and contrasted them with the current site structure. The discrepancies between the two made it very obvious that we were not meeting users' expectations.

Second, we included research data that supported how good visual design and predictability helped instill trust which in turn increases site traffic and usage. This was done to communicate the business impact of the discrepancies we highlighted.

Third, to get area-specific stakeholders, like selling or buying- excited about ways we could address the issues we drafted a high level framework and illustrated how it could solve both user issues within each specific area of the site and issues across the site as a whole. We packed all three deliverables in a PowerPoint deck that told a compelling story ending with tangible ideas on how to solve the issues.

Larry: I think another compelling artifact was a use case walkthrough. We showed what it was like to try to find selling information on the site. The key resources and tools were scattered all over the site -some through the header, some through help, some through the site map, etc. Finally, the predicted Net Present Value (NPV) –which we calculated- of our effort was a valuable deliverable for getting actual approval and booking.

Q: So as designers –what skills did you use to make this possible? Anything particular to design or designing that made things work?

Jen: In order to design any successful solution one needs to understand user needs, content and it's patterns, and business goals. Through research methodologies such as field studies, usability studies, surveys, interviews, etc., a designer is in tune with users needs more than the average employee. We always walk away from a study with a list of key points that need to be addressed. So we have a unique opportunity to see problems first hand.

Larry: I think one key skill that designers are taught is to not jump to solutions. We don't start from a solution and then try to figure out what problem it solves. So many people do that. Designers look at that problems and needs of the users first, then abstract that up a level to get at the "real problems" -not just the reported problems- and then start to think about blue sky solutions to address them, before settling on an actual pragmatic solution.

It is somewhat like the "5 Why's". Don't accept the first reported problem or need. Ask "Why" and ask "Why" again and again until you uncover the "real" problem.

Q: You probably knew this question was coming so here goes… we’ve heard from other designers that defining problems is a way to get further upstream in the product design process –to be more involved in strategy vs. implementation. Was that the case for you guys? What was the result of your work?

Jen: Helping define the problem was very effective at bringing the design team "to the table". In fact, the design team was actually the primary driver of the project. So we were responsible for getting buy-in and resource allocation.

This was a challenge for us, though, since designers were not traditionally owners of projects and it required more effort than usual to gain support from within the company. In the end, we were recognized throughout the company for our efforts, which in turn allowed the design team to lead more projects in the future.

Thanks Jen, Thanks Larry!