Case Studies

I study people and problems to design effortless experiences.

Information Architecture Refresh

Project Overview

The first step in redesigning the website for Harvard’s Center on the Developing Child was fixing the broken information architecture ("IA"). During my time at Fresh Tilled Soil, I conducted user interviews and used tree testing software to improve task success rates by 55%.

Relevant Skills

  • Information Architecture
  • Tree Testing
  • User Experience Auditing
  • User Interviews
“This is much more than a redesign, it’s a complete rethinking of our web presence and I think it sets us up well for the future. We’re really excited about the new site and very appreciative of your desire to get it right for us and our audiences.”
— Senior Leadership Member, Center on The Developing Child

The Harvard Center on the Developing Child

Harvard’s Center on the Developing Child (“The Center”) coordinates researchers and change-makers to drive evidence-based interventions that help children. As such, they wear many hats, serving scientists, philanthropists, intervention program managers, and parents. When I started work on the project, the first challenge was to rethink the site’s confusing information architecture. There was no clear hierarchy to the content, and it didn’t align with the needs of The Center or its constituents.

 
Before: The original Center website made if difficult for users to find what they were looking for. In particular, site content was organized around its format (ie. "Brief" vs. "Multimedia") rather than content itself.

Before: The original Center website made if difficult for users to find what they were looking for. In particular, site content was organized around its format (ie. "Brief" vs. "Multimedia") rather than content itself.

 
Testers using the original IA had a 40% success rate across all tasks. With our new IA, that jumped to 65%, representing a 55% improvement.

Discovering Users and Needs

My team* and I conducted interviews with members of each of the above-listed user groups to better understand their relationships with The Center. Through this research, discussions with the client, and analysis of the site’s visitor analytics, we identified that the three most important types of users were Practitioners (intervention program managers), Big Bettors (philanthropists), and Media (members of the press). Additionally, we isolated three major components of The Center’s mission – facilitating science, implementing interventions, and scaling knowledge – and found that user needs tended to align with these three areas. We sought to use the site’s updated architecture to help clarify its mission. From our research, we generated a list of tasks that users wanted to accomplish when visiting The Center’s website. 

A major challenge in meeting users’ needs with a new IA was the extensive and diverse content library that The Center manages. The Center produces and curates a tremendous amount of content in various forms and on various topics. The old site’s IA was organized around content form (reports, briefs, videos, etc.). The redesign hinged on reorganizing this content to make it easier for users to meet their needs while also implicitly reflecting Center goals. I collaborated with my team to generate a new IA to do just that.

Testing for Success

We were confident that the existing IA was not meeting user needs, but we needed to validate that assumption. We used Optimal Workshop’s tree testing tool to get a quantitative sense of how the existing IA was performing and used this as the benchmark for our design to improve upon. The test is composed of a series of tasks that testers must complete using only the site’s navigation tree. Testers are given the highest level of the tree – the main navigation – and are asked to navigate through the IA to select where they think the solution to the task resides. 

 

An example of tree testing in action (image provided by Optimal Workshop).

 

Tree testing is a great tool for validating complicated IAs because it forces users to make sense of the site’s organization without the bells and whistles of a full-fledged design. If the IA passes tree testing, we can confidently design the rest of the site or application around it. Additionally, having testing benchmarks makes it easy to evaluate future changes. Testers using the original IA had a 40% success rate across all tasks. With our new IA, that jumped to 65%, representing a 55% improvement. 

From IA to Implementation

With the help of my Fresh Tilled Soil colleagues, we moved from redesigned IA to redesigned website. Our redesign makes it easier for users to get what they need and positions The Center to be even more effective in achieving its mission.

 
After: The new site features the updated IA, reflecting user needs and The Center's mission. 

After: The new site features the updated IA, reflecting user needs and The Center's mission. 

 

*Note: I produced this work while I was employed at Fresh Tilled Soil.