Let’s set some context…

When I started at MongoDB in September of 2020 the Education department (where I sat) was in a bit of disarray. Many projects were started but had stalled for one reason or another. The project that follows is one of those stories.

Previous leadership had noted that our navigation problem (detailed below) and how it was not organized and needed to be resolved so that users could use the sites in a more actionable way. However progress on the project entirely stopped before my entering the company.

Im thrilled to report (spoiler) that through my leadership and work I contributed we were not only able to solve the problem but launch this complex company wide project.

The Problem

MongoDB, like so many modern technology companies, grew very fast and as a result had several web properties across their company portfolio. Although this in and of itself is not an issue, what was an issue is that all 4 websites did not look related in any way. The sites had different design systems, different navigation, and there was little or NO way to get from one site to the other.

The biggest impact we though we could make to this problem was to figure out how we can have a universal navigation across all the sites to help resolve some of these issues.

Before state
The video and screenshots below are the best way to see the problem we were looking to solve. Navigating between the 4 sites was almost impossible (even for us internally) as the links were not labeled well, intuitive, nor organized in any reasonable way.

The mongodb.com site simply added links to the new sites/pages in a rapid way, prioritizing speed over usability, which is entirely understandable but couldn’t go on any longer, it had to be addressed.

User research
Our users, when asked in interviews, had no idea that the other sites (additional resources) existed unless they Googled it. In the research section later I will detail more the level of confusion this was causing in a measurable way using “lostness scores.”

My Role and the team

Product Design Lead, overseeing our University, Documentation, and Developer Center sites.

In this project I was responsible for strategy, stakeholder management, the end to end product design process, and company wide communication about the project.

The team - Due to the broad nature of this project it involved MANY teams including: Product design for education (my main team), Marketing, SEO, Visual design, and more.

Core IC team:
Product Design and Research - Erin Essex
Visual Design - Andrew Lee
Lead engineer - Angelo Reale

Focus

Before we got started on the design thinking exercise I wanted the group to agree on where we want to narrow our focus for this project. The reason I felt this step was crucial to agree on up front is that in previous attempts for this project the focus was so far into the future that it wasn’t in fact actionable in a near term way.

This diagram helped us agree on our timeframe and probably of execution.

Method

At MongoDB the company culture really values collaboration, its a value dubbed “Build Together” - and for this project we had to embody that value.

The method I decided to use here is a modified design sprint. We had 5 departments and many stakeholders that could not be narrowed down for a project of this scale.

Design thinking workshops

I love design thinking workshops as a means to get aligned. This project, however, has a ton of stakeholders (as outlined above). Below is the approach I decided to take for this unique situation.

Groups - with a group of 21 employees with very busy calendars I decided to break it into 4 groups, each group went though 2 workshops. Therefore I ran 8 workshops total

Part 1 (4 groups, 2 exercises)

1. Align on goals

2. Open card sort

Part 2 (4 groups, 3 exercises)

1. Lightning demos

2. Sketching

3. Solution drawing

Asynchronous voting

1. Dot votes on drawings

2. One “super vote”

Synthesize results for the group

After all these actives it was critical to share what I had learned, where we were aligned, open questions and clear next steps. You can see what I presented in the deck lined below.

Baseline data

I had the added challenge with this project that we had no tool that was measuring cross site engagement or metrics (a tool like Segment) and therefore sadly I had no traditional way to get baseline data to show improvement. Had I had this I would have used the time on site and the page count metric as a baseline to measure success.

In lieu of that I had to get creative on how we could show impact. Then I remembered using Lostness scores!

“Lostness” pioneered by Tomer Sharon to measure how “lost” users are while trying to complete specific tasks. The idea is that as UX designers when doing information architecture we want to use the clearest terms and navigation groups so that users can find what they need as quickly as possible. However sometimes when we watch users (or are the user) we will bounce round the nav and finally find what we need. What Lostness scores does is count all those attempts and compares it to the ideal path and creates a number that shows how ”lost” someone is.

For our lostness test we identified 8 tasks that were core to the business and had users complete these for our tests.

We ran our current site through the lostness scores and task success rate and here is what the result was:

Ideation

After some of the workshops and card sort testing we moved further into ideation phases. To start I created 4 different prototypes with some of the core takeaways from the workshops and initial testing.

I loved seeing these results because this is one of those occasions where no matter what we do its better than what we have today. So that was a HUGE win. Then it was just a matter of mixing and matching the approaches to have the best outcome for our users.

.

Measure

The images below represent how we were able to leverage our lostness score metric and our task success to show the before and after state of the navigation changes.

I want to call out that I am disappointed that I was not able to show any metrics around time on site or page count as part of the results. We had limitations at the time in that we didn’t have a tool procured at the time that allowed us to measure these. However, I am happy that the lostness scores and task success did show that we positively impacted the experience.

Was the project successful?

Yes. This is a project that was long overdue and was in need of a leader to drive it, and I was happy to be that person.