At the start of this project, I focused on understanding the goals and requirements while reviewing the existing setup. Initially, my role was scoped to simply designing the look of the notification and handing it off to the developer. To ground myself, I reviewed the Slack app already in use for test notifications and the HDL web app to explore the scan-related attributes that could inform the notification design.

However, I noticed the test notification showed very limited information and lacked critical user insights or feedback. Recognizing this gap, I took the initiative to propose expanding the project scope to include a user survey to uncover needs, pain points, and preferences, as well as user testing to validate designs. Once my mentor and my PM approved these goals, I began an in-depth research process to create a more comprehensive and user-focused solution.

To better understand Slack app capabilities and analyze trends in similar tools, I conducted competitive research.

Our HDL Manager app had two primary functions: sending notifications triggered by task completion and connecting to web applications via APIs. I explored the Slack Marketplace and identified five comparable apps — Zapier, Trello, Asana, Jira, and GitHub.

By analyzing these apps, I uncovered common functionalities and visual components in task-driven notifications. This research informed the development of a tailored set of best practices for our app’s notification design.

After competitive research, I developed a detailed user survey to better understand users’ workflows and notification preferences. The survey accounted for the unique workflows of three primary user groups—FC, Madden, and Capture—who managed different parts of the 3D head asset creation process: Raw Scans and Game Resolution Head Scans.

FC handled both scan types independently, Madden ran only Game Resolution scans and relied on Capture for Raw scans, while Capture solely ran Raw scans for Madden. Understanding these distinct workflows was crucial to crafting survey questions that captured diverse user needs.

The user survey had 3 clearly defined goals:

Based on these goals, I came up with 16 unique questions covering users’ role, workflows, Slack notification preferences, and insights about their challenges.

Survey was a huge success, providing crucial insights about workflows and preferences of 9 different users across all 3 primary user groups.

Findings were synthesized using affinity mapping where I organized identified observations, ideas, and direct quotes into 4 categories: Slack Channel & Mentions, Notification Types, Notification Content, and Customization & Control. Then, I presented these insights alongside findings from the competitive research to my team. I created a comprehensive slide deck to highlight user needs, pain points, and best practices, asking for feedback on next steps and aligning the team on priorities.

Based on the findings from the Discover phase, I defined key features for the notification system and created user journey maps to guide the design.

There were a couple of key challenges in defining notification features.

With six possible take statuses (submitted, queued, running, complete, errored, cancelled), different users had varying preferences on which statuses to include in the notifications. So I went back to the survey results which showed "complete" and "errored" statuses as the only statuses that all users wanted to see, so these were the only statuses selected to be included for the notification.

By focusing on the "complete" and "errored" statuses, the design directly aligned with user needs and expectations. This improved the clarity and relevance of notifications, ensuring users only received the most pertinent information.

To address this, I created two versions of notifications for user testing—one with a list of scanned takes and one without, alongside a concise high-level summary.

This approach allowed me to confirm user preferences instead of relying on assumptions.

Based on the survey responses, I developed detailed user journey maps for FC, Madden, and Capture teams. These maps illustrated how each group interacted with HDL from start to finish, emphasizing unique workflows and shared patterns. By mapping touchpoints, pain points, and notification opportunities, I identified when and where notifications would provide the most value. This process informed the notification system's structure, ensuring it aligned seamlessly with workflows, minimized disruptions, and addressed each group's specific needs.

The design phase saw two major iterations, focused on developing a streamlined, user-centered notification system.

I started working on the initial design by leveraging Slack's Block Kit Builder and Figma UI Kits to create accurate, reusable visual components, which helped save time in future iterations. During this period, I focused on understanding Slack UI patterns and playing around with a few parameters, focusing on visual hierarchy, UI principles, and technical feasibility. This allowed me to hit the ground running once the user survey results came in.

The 1st iteration incorporated user insights and defined features from the previous phases. Most importantly, two versions of notifications were developed to confirm whether users wanted to see a list of scanned takes in the notifications. I’d like to dive deeper into some of the key design decisions I made in the 1st iteration.

The notification’s content remained largely unchanged from the initial design, as it already aligned with the users' needs based on survey feedback. Several UI improvements were made to optimize the design however, particularly by reducing the notifications vertical length. This allows for maximum space efficiency for Slack’s limited real estate while showing only the most relevant information.

As previously mentioned, users wanted to see a list of scanned takes in the notification. Now the question was, what’s the best way to present this list? And how does adding the list affect the rest of the design? By analyzing the survey responses, I noticed that most users organized the scanned takes by talent, when communicating the scan results to their stakeholders. As a result, I landed on a design that reflects this workflow, with a collapsible "Breakdown by Talent" section that maximizes space efficiency.

In this Validate phase, I conducted 3 user testing sessions with different primary user groups to validate the first iteration designs.

Users were asked to walk through 2 Figma prototypes with different designs but same flow, which allowed for direct interaction and comparison between two prototype flows.

The flows were structured to progressively introduce complexity, ensuring a clear and focused evaluation process.

This approach balanced clarity and realism, ensuring actionable feedback while minimizing user confusion.

Heading into user testing, I had clear goals defined to confirm users wanted "Breakdown by Talent", validate whether users were satisfied with seeing only the "complete" and "errored" statuses, and clarify further details of user workflows.

User testing revealed critical improvements to the design, such as:

Findings were synthesized using affinity mapping where I organized identified observations, ideas, and direct quotes into clear categories.Then, I presented these insights to my team, again with a comprehensive slidedeck proposing next steps and asking for feedback. Especially when I introduced a new notification level, the team and I had hour-long calls exploring different options until they were eventually convinced that the new notification level was the most optimal solution.

User testing insights presented a new set of challenges.

Previously, notifications were designed to be sent out by import session since that was the only grouping HDL had for takes. However, user testing revealed that each user group had different approaches when submitting takes for scans, making it challenging to find a unified solution.

The approach I took to solve this problem was putting together a diagram of HDL inventory hierarchy and analyzing the desired notification level from each group. As a result, I introduced a new notification level, “batch”, which refers to a group of takes submitted for scans together.

Then, I realized FC and Capture effectively aligned with batch-level notifications, as a batch in Capture’s case equated to an import session. Finally, Madden's preference for notifications by take was considered impractical because users often worked with over 100 takes at a time and sending a notification for each take would overwhelm the Slack channels.

Hence, notifying by batch was adopted as the optimal, scalable solution. This ensured scalability across teams with different workflows and improved information efficiency.

Additionally, while many new ideas came up in user testing, we prioritized the most impactful, low-effort features to ensure timely delivery within the MVP scope.

The iterate phase was focused on implementing key user testing insights into the final design, seeing 2 major design improvements.

Switching to a batch-level summary enhanced clarity by consolidating multiple session-specific notifications into one cohesive update. This approach allowed users to quickly understand the overall status of a group of scans without being overwhelmed by fragmented updates. By focusing on batch-level reporting, the notification system aligned better with user workflows and improved the overall efficiency of information delivery.

The second major design improvement revolved around the updated breakdown section, focused on delivering the most relevant and actionable information. By introducing "Breakdown by Session" and "Errored Takes by Talent," the design provided clarity on which sessions were included in each batch and linked errored takes to their respective pages in the HDL web app, allowing users to prioritize their focus and resolve issues more effectively.

After a comprehensive process of research, analysis, testing, and iterations, a data-driven, user-friendly notification system was put together, addressing distinct user needs and seamlessly integrating into user workflows. It was ready to make character artists’ workflows more efficient, more productive, and more pleasant.

Although the full implementation was not completed during my internship, I contributed by preparing the design for implementation. Using Slack's Block Kit Builder, I translated the final design into ready-to-use code for the backend developer, streamlining the development process and ensuring consistency between design and implementation.

Wrapping up the design process, here are some of my parting thoughts on the project.

As for the key outcomes of the project, I successfully created a seamless, user-centered notification system that streamlined workflows, improving user efficiency. I also validated and iterated on designs based on user feedback, ensuring the system was both practical and effective.

I would suggest next steps for this project to be monitoring usage and collecting feedback from users post-implementation with key success metrics to evaluate the effectiveness of the design and identify pain points. Then, based on feedback, refine the system to address issues such as additional customization needs or workflow adjustments.

Lastly, here are some personal learning and takeaways. Taking initiative to expand the scope of a seemingly simple task, like designing notifications, led to significant improvements and value for the team by considering user workflows, technical feasibility, and potential challenges upfront. I also got to collaborate closely with the team at every point throughout the design process, including regular research sharebacks to the project manager and regular syncs with the developer, ensuring the design was aligned with goals and feasible for implementation.

I'd like to thank my mentor, Haley, our developer, Michael, and our product manager, Allyson, for all the support they've provided for me thorough out this project!