1. Visual imbalance - the look of the platform had an excess amount of negative space that was utilized improperly, while simultaneously cramping content together. Overall, the usage of space was not efficient nor very pleasant to look at.

2. Lack of user interaction - the design is almost exceedingly simple, not guiding the user to specific interactions besides the search bar function. Due to this, it most likely dissuaded many of the user base to not use the Panorama platform very efficiently.

3. Limited features - building upon the last weak point, there are also not many features that the user can engage in. The plainness leads the user to believe that the platform will not solve the original issue that they had.

4. Narrow search function - with the ‘Search Bar’ feature and the ‘Advanced Search’ feature, both are not very captivating and are very ordinary. Other features like a filter or sort are lacking and can definitely be implemented.

Moving forward, I was curious and took a look at other competitors so I could think big and gleaned the best design practices. Internally, Amazon had their own help centers that I could use for inspiration. Here, I presented the Canva Help Center, the Amazon Customer Service center, and the AWS Knowledge Center.

The prevailing concepts I saw repeatedly through different companies is that the content was clear and digestible. You can see through these three sites that there’s a high contrast between the text and background, making everything very easy to read. Commonly accessed topics are placed at the forefront and organized into categories, and a search function is easily found so users can quickly browse for content.

I was still curious about other knowledge bases, so I conducted a competitor analysis, taking a look at seven other companies that had their own knowledge base, ranging from internal AWS and Amazon Help Centers to Spotify Support, Nike Help, and Canva Help Center.

All the stickies at the top are the seven elements proposed for the project scope; keyword search, suggested articles, upvote and downvote capabilities, out-of-date flagging, multiple language availability, requesting information, and displaying analytics. I found that many competitors prioritized the search function and placed it at the top, and asked users if they found the help center to be useful. Another finding was that many sites did not show the last day an article was updated and also lacked the ability to display analytics, like upvotes and downvotes.

With the information I found through the maps, I aimed to hone in on the ‘investigate’ and ‘research’ phases, where both Keisuke and Christine experience feelings of frustration and neutrality. As Keisuke is veteran employee, I believed that integrating a ‘Common Topic Category’ feature would streamline his workflow and minimize his dissent that arises from a slow, inefficient knowledge base. For Christine, I thought that a more focused search function along with transparent user interactivity would help her needs as a new HRSA. This would be implemented through displaying upvotes and downvotes so Christine can see if an article was helpful to other employees at a quick glance.

From all the background research I conducted, I found three main themes:

With mapping, I decided to create customer journey maps to pinpoint the type of scenarios most users use the Knowledge Base in. As a disclaimer, these customer names are falsified and fictional. I created two maps, with Keisuke being an employee with years of experience and Christine being a new hire. I used the observational shadowing of participants showcasing Panorama to create these maps and noted common tasks, emotions, and possible areas of improvement.

With the key themes in mind, I started designing my initial low fidelity wireframes. I created multiple variations with slight differences. All versions have the search bar at the top for ease of access and functionality. I designed a smaller and larger preview to compare whether seeing more articles or more content was preferred.

I also made versions incorporating the ‘Common Topic Category’, displaying four at a time or three at a time. A scrolling component can also be seen as one currently doesn’t exist within the current design library, where a secondary ‘See more’ would be used instead. Finally, there is a mock with an expansion feature as well as a fly out when agents would like to read more information.

This is how the page would look if the whole Knowledge Base was expanded, with a different proposed location for the ‘Suggest New Content’ feature. This design also includes a ‘Last updated date’, and a ‘Menu/options’ feature. If an employee wanted to expand the resource article, it would look like this with a fly out.

Once my low fidelity designs were created, I received numerous feedbacks from the design team I was working with and constructed a final low fidelity design for usability testing. I conducted interviews with five participants running through a lo-fi prototype, where I found these key points: my participants wanted to know more. Users frequently asked how the CTC (Common Topic Category) and Suggested Articles chose what information was to be displayed - which was through a different platform within the project team I was in. That separate platform would essentially gather data from the employee and other staff to consolidate the most frequently ‘clicked on’ articles, and would then display the most popular articles.

I also found that there was confusing verbiage. Four out of five participants did not see the ‘Suggest New Content’ feature and were unsure what that feature even meant. Participants also wondered if the ‘Is this helpful?’ button at the footer was supposed to be for the whole knowledge base or one single article. Users also struggled with the lack of user freedom and would be clicking around looking for a way to go back to the previous page.

Next, I found that the ‘Leave some feedback here’ input box caused uncertainty, with a user stating that “I don’t really understand what the purpose of the feedback is going to be, what’s going to happen with it. It kind of worries me a bit...what are we going to do with those feedbacks?”. Two out of five participants had shared that the copy text feature would not be helpful, considering these resource articles are written for HRSAs and wouldn’t typically be shared with Amazonians needing help. However, I found that my implementation of a clean structure was successful as four out of five users mentioned the aesthetic upgrade of the KB.

I also found the most to least useful features through my usability testing. I had participants rank a total of ten features, 4 features being North Stars that were not in the original project scope. This graph is a depiction of the average rank from 1 to 10, with a higher ranking signifying that the participant believing that the feature would be most useful. Each bar represents a feature, with F&CF meaning Filter and Category Flairs, Sort, CTC meaning Common Topic Category, SGN as Suggest New Content, Favorites, Report, and Leaving Feedback.

Findings show that participants would benefit most from a Filter & Category flair feature, followed by the Common Topic Categories, and a tie between the Sort and Favorites feature. The Filter & Category flair feature along with the Sort are both North Stars, with my results revealing the impact of their implementation. My recommendation would be to add the Filter & Category Flair feature, as well as the Sort feature for P2.

Leveraging my new-founded foundation of research, I constructed my final high fidelity prototype. Here you can see the whole platform, with the knowledge base nested into the right side. Now, I’ll start walking through the Knowledge Base as if I was an employee.

If this was my first time opening up the platform, I’d take a brief glance at everything, and click at the Information icon at the top since it catches my eye first, satiating my curiosity of wanting to learn more. I’ll see that it explains how the Common Topic Categories and Suggested Articles are displayed, then move downward to see the search function. Let’s say that the article I’m looking for isn’t part of the three categories being displayed, I’ll go ahead and use search to find a specific resource. Let’s say this page re-populates with my search term, and I find the article I’m looking for, I’ll go ahead and click on Read more to see more content.

This was helpful to me, so I’ll click the thumbs up, and I’ll use it in the future, so I’ll click the options button, to add to my favorites. Within the options button, I could also change the language in case I'm an HRSA in a different country. On the other hand, let’s say that this article was outdated as there was a recent policy change, I’ll go ahead and add feedback at the bottom detailing that the article is not aligned with current policies, and also click to report an issue. I can click on the back button here that now allows for user control and freedom, or on the home button at the top.

If I can’t find the solution to the case I’m working on, I can click on ‘Request new content’, which has been updated with more concise copy. Now let’s pretend it's a new day and I want to take a look at my old Favorited articles, I'll click on the star. I want to take a look at all these articles with a bigger view, so I'll click the expansion here.