Duration
1 Month (Feb 2019)
Team
Ritu Parekh, Nathalie Rayter, Tyler Stern, Wuyang Wang
My Role
User Interviews, Prototyping (Lo-, Mid, & Hi-Fi screens), & Final Animations
Tools
Figma, Principle
95 Minutes
The average duration for downtime at one of Markel Food Groups' manufacturing plants.
The process to find and replace broken parts is an unnecessary slog. Engineers have to pour through hefty manuals and archaic computer programs to find bin numbers for replacement parts.
A new parts management system that compresses unexpected downtime by streamlining the identification and retrieval of replacement parts.
Parts Pantry is expected to cut downtime to 32 minutes, saving Markel's customers time and money.
Scanning Part Info
Native mobile application capabilities allow for haptic feedback and scanning QR codes attached to machinery.
Part Info at a Glance
Engineers can locate parts based on its location within the equipment and preview information about the part before reporting it as damaged -- logging that damage data with the parts maintenance system.
Share Location
Parts Pantry works within existing walkie-talkie user flows, allowing engineers to share a requested part's location in the storage room with team members, expediting the repair process.
Order Parts
Parts Pantry allows engineers to order replacement parts while the application passively collects data on damage and order history.
Markel’s AMF Bakery, a subsidiary of the Markel Food Group, is a product-service system, providing physical bakery equipment as product and system design and support as services.
The Parts & Service team provides technical assistance and trainings, delivers responsive parts, and comprehensive equipment evaluations. Their goal is to minimize downtime for their customers.
Our generative research took the form of researching Markel and AMF online, but our most fruitful method was interviewing technicians at food manufacturing plants.
In these interviews, we learned about the general work flow on the plant floor, what happens during downtime (when the machinery is stopped for repairs), and how most of the infrastructure in place to find replacement parts is clunky, inefficient, and frustrating.
Controls Engineer
The Haskell Company
"I am happiest when I have problems I can solve and the resources to do so."
Maintenance Technician
Frito Lay
Machines that don't work as expected become "personal challenges."
From our generative research, we created three personas to help us craft a narrative for our users and better communicate breakpoints.
From our interviews and personas, we felt as though Bill's experience with machinery offered the most fruitful design opportunity. We modeled his persona's encounter with downtime in a journey map to visualize the breakdowns identified in our research.
Reversed Assumptions
Flipped assumptions from generative research to unearth new services and ideas
Collaborative Sketching
Drew rough wireframes of a maintenance tool using the "Yes, and..." principle
20 Questions
Generated 20 questions related to one fact: downtime costs too much money
Lo-Fi Prototype
Our first low-fidelity wireframes depict a facility management native app. It aims to reduce Markel customers’ production line downtime by providing their maintenance technicians with a platform to do their jobs more effectively.
The home screen of the app featured system status reports, work logs from previous shifts, and capabilities of filing reports and ordering replacement parts.
During our ideation phase, we landed on the idea of an augmented-reality feature by which bakery facility maintenance teams could easily visualize the the system status as they view the machinery through the camera on their device.
However, we decided to nix this feature of the system status app; the added value that AR annotation would create does not justify the amount of development effort and processing time it would require to be a successful feature.
Instead, we planned on using a system map with location accessibility.
Throughout our design process, we considered what our interview subject Bob the maintenance technician said:
“I am happiest when I have problems I can solve and the resources to do so.”
Through this process, we endeavored to design an app that would equip people such as Bob to do their jobs well by providing instant access to part information and manuals.
During our in-class critique, we were told that our scope was too wide and solution too complex. We needed to figure out how to narrow and focus on just one way to decrease downtime.
Our narrowly focused mid-fidelity screens aimed to reduce Markel customers’ downtime by decreasing the amount of time it takes a maintenance crew to locate replacement parts.
The app is scoped to only include service bearings, as these are among the most problematic parts in industrial food systems.
We were still attempting to do a lot by integrating parts installation instructions and detailed part information into the features. We needed to better consider the needs of the technicians who would be using the app:
Do the technicians have the authority to order a new part?
Would the installation guide be useful to them?
Reflecting on this feedback, we decided to narrow our focus to just one part of this flow and use that to to decrease downtime in the bakery.
In our next iteration, we aimed for this app to serve three main functions:
In this version, a technician would scan a QR code on the section of the production line they need to repair. The app would return a machine schematic for the tech to identify the part.
The app then displays the part. After engineer reports it as broken, it provides the location of the part.
This iteration of the app also allows for sharing the inventory location with another technician who could then be radioed to retrieve it.
Finally, we switched order parts to a simple order request that would be pushed to the engineer's manager for approval.
Our micro-interactions come to the forefront in these final screens. Confirmation screens give joyful, donut-based feedback, while proceeding through the screens slides with constant forward motion to mirror the mindset of the industrial bakery.
Scan QR codes, provide haptic feedback, and offer real-time location data
Scan QR codes, provide haptic feedback, and offer real-time location data
Big, easy-to-push buttons for one-handed work situations
Passively collects data about product and part lifespan to inform future work.
We greatly benefited from a strong research foundation in this process. After doing background research and interviewing stakeholders, we were able to identify specific breakdowns in the current industrial baking system. This helped ground our ideas in a concrete way, helping us ideate more relevant possibilities. Critique steered us in the direction of scoping, and we increasingly narrowed down to one specific goal: compressing unplanned downtime through part identification, and one specific part: bearings. This helped us to be more focused and came up with more feasible solutions.