Why user-centered design research is time-consuming and totally worth it

Three of state government's biggest design proponents told a conference audience how and why they should consider a new approach to their digital government solutions.

“This is usually when I throw the phone across the room.”

Those words were spoken by tester of one of the State of Georgia’s mobile apps. It’s just the kind of scenario the state hoped to find during testing so it wouldn’t happen with a real user. And it’s one way to ensure that government services have intuitive user-centric interfaces like those offered by the private sector. 

These days, that kind of user-centered design is high on the list of skills that many state agencies want to improve. “Government must keep pace with the Amazons of the world” is a common refrain in the state IT community. 

Attendees got a little help with it Tuesday in Baltimore at the National Association of State Chief Information Officers (NASCIO) annual conference, when three of the leading user-centered design experts from state government offered a few tips.


You might not need a mobile app.

Nikhil Deshpande, chief digital officer for Digital Services Georgia (DSG), walked an audience through how the state uses user-centered testing to design the state’s websites and applications. Deshpande shared his first lesson in an anecdote in which a customer agency approached DSG and said they need an app.

“That’s when we took a pause,” Deshpande said, “stared them in the eyes and said, ‘Do you?’”

Rather than starting with the solution, it’s best to examine the problem first, Deshpande said, and ensure that a mobile app is the best solution, rather than a responsive mobile-friendly website, or maybe some other solution.

That idea was cousins of another in the same category that Deshpande pointed out to the audience, which is that the way users interact with a government’s digital presence may be different than imagined. The actual way a state’s websites and apps are being used should guide development, he said.


Spend your time on the right things.

Deshpande said that when the state designed its main website in 2002, it worked from an assumption that everyone would start at, a central portal, and then be routed to the correct website on one of the many branches of the state government’s expansive tree of services. It sounds reasonable enough, but through testing — which in Georgia includes user testing and video recordings of what the user is doing and thinking as he tries the latest version of a digital service under development — the state found that more than 60 percent of users were arriving to webpages through a Google search, and very few were using their interface as intended.

There are a lot of tools that can help governments spot ideas like this and others, and Holly St. Clair, the chief digital officer for Massachusetts, showed the audience how her state used a few of them to redesign the state’s website starting about 18 months ago when she was asked by Gov. Charlie Baker to create one that was both “data-driven and constituent-focused.”

Don’t get in the way of helping the user.

In its zeal to serve the citizens, the state had provided some 250,000 pages of documents online, but only 10 percent of the content drove 89 percent of their traffic, St. Clair said, which means that most of the content was getting in the way of the things people were looking for.


She said her team used tools like Google Analytics to conduct cluster analysis on the types of topics that people were interested in. They used Chalkmark to gather first impressions on designs and Treejack to visualize the navigation paths that test users took, allowing her agency to adjust wording or menu hierarchies when users kept taking wrong turns. Crazyegg provided the team heat maps that showed where people were clicking. What they found, she said, was that their previous attempts to help the user were often doing more harm than good.

And that’s the importance of user research and testing, St. Clair said — it’s not possible to know how something should be designed from pure logic alone because users will reveal all sorts of behaviors and interpretations that cannot be anticipated. She acknowledged that this is a challenge in government, where leadership will want something in-hand after all the money on research and design has been spent, but nothing has been produced.

Get something usable fast and iterate.

Her recommendation: Get something something that is functional into the hands of users fast. An iterative design and development approach can always be improved upon, and the users will ensure that process is informed by their wants and needs.

Karen Geduldig, deputy chief information officer for the New York State Office of Information Technology, agreed that while there can be steep up-front costs associated with data-backed user-centric design, there is an upside that the same skeptical executives can appreciate.


Though time-consuming, user research is also reusable.

“We now have a framework we can leverage across all the agencies we support,” Geduldig said of the user research that produced a mobile app created for the New York State Department of Agriculture and Markets’ plant inspectors.

The app was created as a replacement to the paper-based process that had been considered highly inefficient.

“The Department of Agriculture does many inspections, but with plants it was a very paper-based, data-entry driven process,” she said. “They have books all across their backseats and they would go out into the field and bring their encyclopedias of plants to make sure they were healthy and prolific.”

The new app autofills fields and saves inspectors time in the field, she said.


User research pays off.

Perhaps most importantly, Geduldig said, when the research is done and the app or website works well, the state is left with a tool that is effective and can prove well worth the investment that was required to bring it to life.

“Just the sugar on top of the cookie, we have increased the productivity of Ag and Markets,” she said. “It used to take about a day to a day and a half to do one to two inspections. Now our plant inspectors are doing three a day. And that is particularly important when we look at resource constraints across government.”

Latest Podcasts