BizdevIQ — A UX Case Study

Designing a Web App to Connect Data Science, Machine Learning and AI Students with Companies for School Projects

Image for post
Image for post
A screenshot of the new, proposed BizdevIQ homepage.

Introduction

As teams are working to adapt to remote work in the current “new reality” that COVID-19 has created, half way through the project, my team had to quickly adjust to working remotely and still deliver to our client by the same end date for our contract. BiddevIQ, a tech startup, hired my team of 3 UX designers to work on a 3-week design sprint, but more on that in a bit. Throughout the 3 weeks, my team found ways to research, present, white board, affinity map, sketch, and collaborate remotely and with ease, in order to present a high quality deliverables to our client.

Project Overview

Objective:

Our objective for this project was to design a website platform for BizdevIQ that would support it in fulfilling its mission to connect students with companies and provide them with artificial intelligence, machine learning, and data science projects.

BizdevIQ’s goal is to help students build their portfolio before they graduate so they are prepared for the job market after their program.

Users: Machine Learning (ML), Artificial Intelligence (AI), and Data Science (DS) Students

Timeline: 3-Week Design Sprint

Team + My Role: I was on a team of 3 UX/UI Designers. I contributed mainly as the lead UX researcher, and the project lead, creating timelines, calling team stand ups, and ensuring deliverables would be ready by our agreed upon delivery dates.

Tools:

  • Figma
  • Pen + Paper
  • Google Slides, Google Docs
  • Slack
  • Zoom (This became even more important 1/2 way through the project)

Scope Decisions:

Because it was a 3 week design sprint, we spoke with our client and agreed to focus on designing the Minimum Viable Product for students. Doing user research on the needs, goals, behaviors, and pain points for the other users of this site, i.e. Companies and Universities, would have to come in a later sprint.

Research

Recruiting Users to Interview

After agreeing upon our Statement of Work, our client gave us a list of current or past users to reach out to for user interviews. After sending an initial round of emails and follow up emails, my team realized that we’d likely have to do some recruiting of our own for users.

Image for post
Image for post
Screener Survey Used to Find Users to Interview

In order to ensure that we found the right users for our research, we created a screener survey on Google Forms that screened for users who were either current or recently graduated masters level students focusing on Data Science, Machine Learning, or AI. When we did not have much luck with that narrow focus, we expanded our user research recruiting to include General Assembly Data Science Immersive students, as well.

User Interviews

Ultimately we were able to find a total of 7 students or recent grads who met our criteria to interview.

As mentioned above, for this project we decided to use the methodology of user interviews because they would provide our team with valuable insights into the mental model of BizdevIQ’s primary users — i.e. students. We chose to do this because it became clear in our kick-off meeting with our client, that BizdevIQ lacked substantial user insights and data on it’s potential users and their needs, goals, behaviors, and pain points.

The make up of the 7 students we interviewed was:

  • 1 current AI Masters student
  • 3 recent grads of AI, ML and DS masters programs
  • 3 General Assembly Data Science Immersive students

In our interviews, we aimed to gain insights into DS, ML and AI students’ goals, needs, pain points, behaviors, values, perceptions, and experiences around building their portfolios before graduation.

Contextual Interviews

In addition to conducting student user interviews, my team also did a few contextual interviews to get a better understanding of the problem space. We interviewed a professor at Fordham, who has worked with BizdevIQ in the past, and he gave us insight into the current reality of what it means to be an AI, DS, or ML Masters Student at Fordham University, and what the expectations are from the University’s side.

We also had the opportunity to interview a representative from BMW Canada, who had worked with BizdevIQ in the past year, and had created a data set that he made available to students at Queens University in Toronto to work on for one of their school projects. In speaking with the company representative, it became clearer and clearer that BizdevIQ would need to do further research into the company side and the value proposition for companies to work with them. In speaking with the representative, the amount of time and permissions it took to create a project that students could work on did not seem to be worth the effort for a large company like BMW, and the main motivator for this representative was that he had graduated from the same program 2 years earlier and wanted to make real world data projects available to students in that program. When we asked the representative if a company like BMW would consider hiring students to work on projects like this — he said that ultimately they were more likely to hire a large consulting firm to do the work that needed to get done, over students, because of the quality guarantee.

As my team delved further into our user interviews and contextual interviews, it became clear that more research needed to be done regarding what kind of companies would sign up for this platform and be willing to do the work of making their data available to students. That said, because we only had 2 more weeks left in our sprint, my team continued to focus on the needs of the students, the primary users, and made recommendations to our client about further research they should conduct.

Research Synthesis

Image for post
Image for post
Research Synthesis: Using the Process of Affinity Mapping to find Common User Insights

Insights from Affinity Mapping

In order to synthesize all the data points we uncovered in our user interviews, my team chose the method of Affinity Mapping in order to quickly discover and distill user insights that would drive our design process. We ultimately settled on a total of 12 insight groupings, and for the purpose of this design sprint, we chose to focus on the following 4 main insights from our user research.

Key Insights:

  • I need expert guidance
  • I want to work with clients
  • I need soft skills to stand out
  • I network to get work

Persona

Using the insights gained from affinity mapping, my team developed a user persona, who would embody the needs, behaviors, goals, and pain points of BizdevIQs primary user.

Meet Sofia, a 29 year old masters level graduate student studying Data Science in New York City.

Image for post
Image for post

Sofia’s Journey Map

Now that you’ve met Sofia, let’s look at a recent experience she had working on a group project for one of her DS classes. …

Image for post
Image for post
Sofia’s Journey Map: Her Experience receiving a group assignment in class, and how it turns out.

As we examined Sofia’s mood throughout her journey, we identified two main areas where it dipped, that we have selected to focus on as our areas of opportunity for this design sprint.

Image for post
Image for post
Area’s of opportunity we identified in Sofia’s Journey Map.

Problem Statment

Now that we had a clear picture of how who Sofia is, what her needs, behaviors, goals and pain points are, my team created a problem statement that would drive our design process.

DS, ML and AI students go through very technical programs at university to add to their skill set, but there is a gap in their training/experience when it comes to applying what they have learned to a job in the field.

Sofia is midway through her Master’s program in Data Science and she is trying to make sure she has what it takes to be job ready when she graduates.

How might we help Sofia showcase her unique value to employers while providing her with the support and guidance she needs to land her dream job?

Platform Choice

Based off data points collected from our user research, we decided to focus designing a desktop site for this sprint, given that 7 out of 7 of our interviewees worked mainly on their desktops.

Moscow Map

To help us prioritize what features to focus on for this sprint, we used the process of MoSCoW mapping where we determined which features the Minimum Viable Product (MVP) web app must have, should have, could have, and won’t have for this specific design sprint.

In this iteration, we chose the must have features that related directly to the areas of opportunity for Sofia that we found in our research synthesis.

Image for post
Image for post
Our Team’s MoSCoW Map for the MVP of this Design Sprint
Image for post
Image for post
Design Studio — Rapidly sketching out ideas for what the MVP of this web app would look like

Design

Design Studio

With our problem statement clearly identified, my team moved forward with holding a series of 7 design studios where we quickly sketched out and converged on layout ideas for what the BizdevIQ platform would look like for students.

We wanted to include our client in the Design Studio as well, so that we could ensure we included his insights and knowledge into our designs. Because of growing travel restrictions due to COVID, we hosted a virtual design studio with our client using Zoom, a white board, and paper and pencil. Each of us individually sketched out our ideas for a specific page on either a white board or paper, and the pitched our idea to the team, we then critiqued each other’s sketches, did another round of individual sketching and pitching, until we then brought all the ideas into one final design.

We found that although our client was not physically in the room with us, we were still able to hold a successful and engaging design studio, and we gained important insights into content that was required to build out the project brief sample page.

Mid-Fidelity Prototype Features

Image for post
Image for post

We chose to keep much of what is currently on BizdevIQ’s home page. We cleaned up the website footer, the top navigation, and we added a section at the bottom of the home page that includes information about what companies and institutions use BizdevIQ.

Image for post
Image for post

For the project results page we focused on helping Sofia quickly find and filter for projects that her group could apply to work on based on the requirements set out by their professor.

Image for post
Image for post

Once Sofia clicks on a project to view more, she is taken to a project brief page, where she can read more information about the project, what the objective is, who the subject matter expert is who is assigned, and what kind of technologies she’ll get to work with for this project.

Image for post
Image for post

We also discovered in our user research that Sofia wants to and needs to develop her soft skills and wants advices about how to stand out in her job search. Therefore, we recommended that BizdevIQ also offer a Mentor Section on their site.

Image for post
Image for post

Sofia can search for a mentor that specializes in her area of interest, or has skills or experience that would help her hone particular skills she wants to work on to help her stand out in the job search. On each mentor page, Sofia can send a message to that mentor and request to connect.

Mid-Fidelity Usability Testing

Our next task was to do usability testing to check for the functionality of our design before taking it up to a higher fidelity.

We did usability tests with 5 users, based on the Norman Nielsen Group’s recommendation of only needing 5 users to test on.

We gave our users the following scenario and tasks:

SCENARIO
You are pursuing a Masters Degree in DS/ML or AI. For your entire program you have been assigned to a team for all group projects. You and your team have to accomplish a project of your choice, and it is up to you to choose the data that you will work with for the project.

You and your team have joined BizDevIQ, a platform which allows ML, DS and AI graduate students to connect with companies, who have data projects for students to work on for their school projects.

TASK 1
Log in to your BizDevIQ account. Find a project for your team and ask to join that project.

TASK 2

Your program is focused on Predictive Analytics, find a mentor who specialize in your area of interest and ask to meet them.

Mid-Fidelity Usability Test Results

5 out of 5 of our users successfully completed both tasks that we gave them. That said, 2 out of 5 users ignored or did not see the filter function on our results page, so we made sure to try and make the filtering function more visible in the hi-fidelity mock up.

Image for post
Image for post
Image for post
Image for post

Hi-Fidelity Prototype Walk Through

Click play to see a walk through of our hi-fidelity prototype for BizdevIQ

Hi-Fidelity Usability Testing

For this round of user testing, we did it all remotely via zoom and screen sharing. We aimed to test with users who were not available when we were holding user interviews. The information we gained doing our hi-fidelity user testing with these target users was invaluable, and left us with clear next steps and recommendations to offer to our client for how to proceed with further refining the website for students.

Here are a few insights and issues we identified in our usability testing:

Image for post
Image for post
Image for post
Image for post
Image for post
Image for post

Deliver

Presentation

Two weeks into our social distancing, and at the end of our 3 week sprint, my team presented to our client the results of our work. Although we would have preferred to present in person, my team deftly presented our slide deck to our client, and afterwards we had an hour-long conversation where our client asked us our recommendations and further insights into what we discovered. All in all, our team delivered a high quality product on time and despite the seeming issues that could have come up due to switching to remote work because of COVID-19.

Recommendations + Next Steps

The following are some of the recommendations and next steps we made to our client for this project.

Research Recommendations

  • Conduct more research regarding the possibility of offering paid vs. unpaid projects for students
  • Conduct in-depth user research into the other main users of the site including: Universities, Subject Matter Experts, and Companies

Design Recommendations

  • Work with a UX writer and/or marketing/branding expert on the homepage content and messaging
  • Work with subject experts to streamline the information on project/ mentor brief pages

Next Steps

  • Explore adding a feature for “sharing” projects with teammates on the site
  • Explore different layout on projects/ mentors results pages (like hover states & card expanding)
  • Building out the team pages, company pages, resource pages, events pages, etc.
  • Further usability testing on the updates made

Key Learnings

  • As long as you have a clear plan, use tools such as Zoom, Slack, and Figma, it is completely possible to do a UX design sprint remotely every step of the way and still delivery high quality insights and designs.
  • Furthermore, because we had to do much of our work remotely, we were able to have access to users and students who we may not have been able to engage with if we focused just on in person user interviews and usability tests.
  • Therefore, we gained important information and insights that we might have missed out on due to the time constraints and scope of the project if we didn’t do everything remotely.

Written by

A UX Designer Iterating her Way toward a Better Life.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store