Category Finder
A tool that maps users to the most relevant software category
’Category Finder’ won two 1st place awards (Buyer Experiential Impact and Popular Vote)
Role
Project Lead / Lead Product Designer
Responsibility
Manage the timeline and project, clearly relay idea to the team, keep team aligned on goal, create wireframes and hi-fi mock-ups, edit the front end with CSS and HTML, and present to stakeholders and company
Duration
3 days (Oct 20-22, 2020)
Problem
Buyers often need new software to solve new problems for their business, however don’t know what to call the software they need. Unfortunately, Capterra.com is structured so that buyers need to know which software category they need in order to discover software.
User & Audience
Top of the funnel users who are actively looking for software and do not know what category they belong in
Objective
Help users who don’t know what software category they belong in.
Solution
Use the information provided by our users to provide a list of relevant software categories and products.
Scope & Constraints
This was during our first remote-only Hackathon. We also only had 36 hours to work on this project and most team members were in different time zones.
Step 1 | Choose a Project
In countless user interview recordings, we’ve seen users struggle to discover software products because they don’t know what category the software belongs in. The way Capterra.com is structured makes it very difficult for users to find software categories by their needs. You need to know your category to get to a list of software products.
I also noticed that users felt like they found a relevant category when they found specific features and keywords in a software product’s description. I wanted to leverage this information, as well as natural language processing (NLP), to get users to the right software category.
step 2 | Assemble a team
Starting two weeks prior to Hackathon, I recruited people for my team. Our team comprised of four engineers, one content writer (with a background in informational technology), one brand director, and one UX researcher. I wanted engineers (Shivam, Shashank, Topher, & Rohit) because I wanted to learn more about various NLP technologies (Google DialogFlow, Amazon Lex, IBM Watson) and create an interactive product by the end of Hackathon. I recruited Amanda, who has a background in informational technology, to see how we can leverage different available data to surface relevant information to our users. I brought on Will, an UX researcher because he’s also seen many videos regarding users struggling to find software on Capterra.com. I thought Zev, the brand director, could talk about how he communicates with stakeholders and what high-level strategies are being talked about with leadership.
step 3 | Align on a goal and MVP
It’s important to make sure everyone is aligned on the same goal and understands the user we’re trying to help. We want to help users who are at the top of the funnel and are actively looking for software. These users do not know what category the software belongs in but they have a general idea of their needs.
Using Slack and Microsoft Teams, we met and each member helped fill in knowledge gaps. For example, the engineers did a great job of explaining natural language processing and how we might be able to leverage that for the project. Also, all of us wanted to have a working prototype to show at the end of Hackathon. In order for this to happen, engineers needed design deliverables, such as mock-ups and user flows, and data to feed the feature. After every meeting, I made sure to make a list of action items to follow-up on.
step 4 | Mock-up Creation and Data collection
Once everyone was on the same page about what we wanted to create and how we wanted to address this problem, I created user flows and mock-ups. I wanted engineers to have a thorough picture of what the features should look and behave like, especially since we would be working in different time zones. I also reached out to various teams, such as Catalog, to get access to any relevant data (category names, list of features associated with categories, relevance of various features).
This is the starting screen of ‘Category Finder.’ I wanted to use copy that empathized with the user’s struggle and provided a solution to alleviate that struggle.
When a user clicks on “Get Started,” they will be brought to this screen. It asks for industry, company size, and number of users. This information helps us filter out any one of our 700+ categories that aren’t the right size and too niche.
After a user has completed all three fields, they can continue to the needs section. Here, users can type in what they need their software to do. This is where we wanted to leverage NLP, synonym rings, elastic search, and intent-based relevance to dynamically surface queries that may be closely related to whatever the user is typing. As more and more users use this product, the feature would make connections and learn what the user means to type and look for.
Before we have that data, we can also feed ‘Category Finder’ with data about the categories, its features, and a feature’s relative relevance to that specific category. For example, our features are labeled as core, common, optional, and differentiator depending on how relevant that feature is to that category.
Also once a user adds at least one need, ‘Category Finder’ will dynamically provide other relevant features. I’ve added this functionality because I’ve often seen users struggle to keyword match and come up with feature names on the spot.
Once a user adds all their needs, they can proceed to the next screen which is the results page.
Why show multiple categories?
In previous user interviews, we’ve heard that users liked having options as long as it wasn’t an overwhelming number. Users like feeling like they have options.
Why did I include certain information?
In previous user interviews, we’ve learned that users don’t really understand what’s the difference between various categories. Therefore I thought it’d be great to include a short description, benefits, and typical features of that certain category to provide context.
How would you recommend products?
All the recommended products should include the features/needs they’ve indicated in the previous screen. The products should also be highly rated with large number of reviews. We’ve seen users really value this piece of information about the software product because it helps build trust.
Why include a compare CTA?
Through previous user interviews, we’ve learned that comparison is a very valued tool for evaluating software and the natural next step after discovering software. I wanted to facilitate the user’s movement through the software buying journey.
step 5 | Live Prototype & Presentation
The interactive prototype
During Hackathon, the engineers worked on the back-end, while I worked on the presentation. Once the engineers were done with creating the framework, I edited the interactive prototype with HTML and CSS to make it more aligned with our design system. This was important to me because I didn’t want the inconsistent use of space, color, typography, and text alignment to distract people from focusing on the value and functionality.
If you want to watch the presentation that was shared during Hackathon, watch the video here.
If you want to explore the interactive prototype, click here.
After Hackathon...
Each group presented their project to the leadership team of Gartner Digital Markets. At the end, the leadership team and our peers picked winners and honorable mentions for these following categories:
Business Impact Award (in terms of revenue)
Vendor Experiential Impact Award
Buyer Experiential Impact Award
Internal Experiential Impact Award
Moonshot Award (most innovative)
Popular Vote (picked by peers)
Out of those categories, my project won first place for both - Buyer Experiential Impact Award and Popular Vote Award! This was really meaningful because winning those two awards meant, not only the leadership team, but also our peers thought our ‘Category Finder’ would be most impactful to our users.
As first place winners of Hackathon, we had the opportunity to present our projects to Gene Hall, the CEO of Gartner. In terms of next steps, I am working with our product team to make this feature available on Capterra.com for 2021.