Barnes & Noble Education

An online learning platform for every student

The Brief

Ah, Bartleby! Ah, humanity!

—Herman Melville, Bartleby the Scrivener

Barnes & Noble Education serves more than 6 million college students across the US. They wanted to expand into the online learning space, and hired us to build their education platform, bartleby.

My Role

I lead the design process, kicking off the project by facilitating design workshops to identify problems and find opportunities. I worked with co-designer Matthew Famularo to design the product, translate the brand identity into a design system. I also worked closely with engineering on implementing prototypes into shippable product.

The Team

Will Denton—Product Designer
Grace Pelling—Product Manager
Matthew Famularo—Product Designer
Phil Johnson—Product Manager
Adam Pash—Engineer
Nick Beattie—Engineer


Understanding user needs and competitive advantage

We kicked off the project with a three day workshop, working with Barnes & Noble Education executives to clearly define our shared assumptions, user needs, and business goals. We interviewed 22 students across a variety of fields of study, degrees, and income brackets. Finance, Business, Psychology, and Education had the most overlap between students’ areas of study.

Competitive landscape—what brands are students using?

1. Chegg
2. Course Hero
3. Quizlet
4. Youtube

How do they find out about these products?

1. Word of mouth (Peers)
2. Word of mouth (Professors)
3. Web search (Google)
4. Marketing by the service providers

Pain Point: Many students do not have a clear understanding of the different products and services Chegg has to offer, particularly “textbook solutions” and “Q&A”.

Mental models—how do students study?

We had two distinct mental models when it came to how students sought solutions to their homework. One wanted to find the answer immediately, while the other wanted to see step-by-step how to approach the answer. Students would operate in both of those models, depending on the difficulty of the homework, and the time of day (or night) they were doing their homework.

“Late at night, early in the morning, usually when there's a pressing deadline, like I need to grab this answer right now. …More than likely, it’s a last minute pressing kind of thing..”
“Often times you’re not getting the entire problem wrong. …it could be just one simple number wrong and you’re screwing up your entire problem, …So, I definitely like that it breaks it down into different steps.”

Our three biggest opportunities to serve user needs and provide a competitive advantage were:

1. Make textbooks and solutions easy to find.
2. Design solutions that can get to the answer quickly, and slowly.


A more efficient IA

Rather than directly mimic their biggest competitor, we chose to remove the concept of a solution manual from the sitemap. Students have textbooks and they need answers.

A simplified sitemap comparison between bartleby and their competitor, Chegg.

Searching multiple content types

Users can search for both textbooks and the solutions within those textbooks, in one search experience. Working closely with the clients’ back-end engineers I proposed we rank textbooks and solution results to the search query, and return the most relevant results with elastic search results.

elastic search results, and the search result page.

Evens and odds, linear and non-linear

With a clear path to the textbook paved, and a solid breadcrumb trail for SEO, we still had to solve the problem of working through multiple problems in a text book. The students we interviewed often mentioned their homework would be a sequential set of problems from one chapter, multiple chapters or even a semi-sequential set of problems like "evens or odds". It became clear that students needed a clear way for linear and non-linear navigation. This focused form of navigation informed the structure of the solution page.

Product design informs brand design

The product design team collaborated with bartleby's branding agency to develop the brand identity. I outlined what the product needed from the brand which included a color palette that met accessibility standards, and typefaces which could include Math ML and LaTex glyphs. We complimented Source Serif, the workhorse body copy typeface chosen by the brand designers, with Cambria Math, a comparable yet slightly different serif that could render complex mathematic equations in Math ML.

All the elements of the identity system were informed by product needs, improving the end-user experience and bridging the gap between brand and product.

Snapshots from the design system

Scaling the MVP for new features

After launching an MVP in 5 months, we reexamined our original hypothesis and site map to see how we improve. We discovered that textbooks were a proxy for what students wanted; which was to find precise answers, and also broad concepts related to their learning.

The site's focus shifted from finding just your textbook, to discovering content that is relevant to you, and saving that content.

An expanded sitemap, illustrating the subject page as the hub for content discovery
The dashboard, navigation and study guides implemented post-MVP

From vendor to trusted partner

I spent 11 months working closely with the Barnes and Noble Education Team. From a project code name and a few back-end diagrams, my team built bartleby. A product with a brand promise at heart – to be integral to a student’s academic life. Our client came to trust us because we took on the design problem. We challenged their assumptions for the good of the product. And in turn we had more to contribute to the product.

What did I learn? SEO is not sorcery.

I knew this site needed to be easily found on any search engine. Organic search was how we envisioned most students finding bartleby. But I had to learn what needed to happen to ensure it could be easily found. And more importantly, I learned how to design a user-centered experience with the constraints of what a crawler needs to see.

Where could I have done better? Push back against personas.

In the beginning of the project we built user personas. While our intent was to better understand user needs, The persona exercise quickly became a game of projecting stereotypical narratives onto a demographic dataset. In my opinion, I don’t think personas are terribly helpful in defining design problems. There are other tools that are more effective, that do not reinforce stereotypes about the wants, needs, and concerns of entire groups of people.

I would love to show you how I approached the new features, but it's a little too long for one case study. Shoot me an email and we can continue the conversation.

More Work