Case Study
Convoy Operations Center
Convoy was the leading digital freight network, using automation and machine learning to connect shippers and carriers efficiently and minimizing waste.
I designed an operational workflow experience that allowed Convoy's team to schedule, track, manage, and troubleshoot tens of thousands of live shipments traveling all across the country.
OVERVIEW
Objectives
Convoy fields a robust operations team that specializes in multiple functions. For the team to work at the pace and complexity needed to support live shipments, the initial workflows often relied on managing dozens of spreadsheets, Slack channels, and emails for team members to communicate, coordinate, and execute successful shipments.
Our objective was to develop and implement an evolved operations experience where work could be accurately tracked and completed amongst a team of mixed specialties. Working with operations managers and product, we agreed that not only did we want to improve the team’s speed, accuracy, and ability to collaborate, but to also provide a guided experience where new operations team members could onboard and learn quickly.
A secondary goal was to reduce and eliminate the number of third party tools we were using. An operations task could sometimes require someone to reference multiple documents, chat messages, and other resources scattered amongst multiple applications.
My Role
My role was the senior designer and researcher, working with a team of 1x product manager and 8x engineers. Tasks included:
Design research into the day of the various operations teams, from generative research into opportunities to task observation
Working with operations managers to identify areas of inefficiency and improving the tooling to improve time on task in the product
Collaborating with support staff on integrating and standardizing processes into the product that previously relied on training, external tools, and tribal knowledge
Designing the interface and workflows
Working closely with engineering and product for implementation and delivery
Exploration
Initial Research
Early observation yielded several opportunities to reduce the number of steps, tools and tasks a person needed to take to complete their work. Through interviews and observation for example, I identified users having to access up to 10 different tools in a given session of completing a task - this included Slack, email, Google docs, a customer's website, referencing paper documents, contacting a driver through a texting tool.
This created further challenges for new team members who had to onboard to a scattered workflow full of institutional knowledge.
I also worked with many operational teams to understand and map out how particular tasks moved through different teams and areas of expertise, and the unique workflows to each role.
This was done through a combination of interviews, live in-person work sessions, and even participating myself in the training and operational programs.
IDEATION
Detangling Everything Everywhere All at Once
Because of the sheer number of variations tasks can take, it was important to first work with teams to standardize operational practices. This included documenting and systematizing procedures that could be wholly unique. For example, a supplier for Starbucks might require you to schedule a delivery through a specific website (which only works on Internet Explorer 9). Then, Starbucks requires you to call a specific person for stores that are located in East Washington to confirm delivery instructions.
Beyond working with customer systems, our own processes heavily relied on third party tools. Hundreds of different Slack conversations contained mission-critical information hidden away in direct messages. Thousands of operational procedures contained in binders or Google documents.
Design
Bringing it all together
With the primary goal of bringing all these disparate tools together, I wanted to ensure that these were not isolated features, but rather modules that could be arranged in the appropriate workflow. To achieve this, I created a queue based task system that would both prioritize and summarize key data and actions into flexible cards.
In the backend, teams took the process documents, knowledge, and procedures and input them into an ingestion dashboard that would allow the appropriate information to be surfaced on the appropriate tasks.
To bring the right sequence of events together, I worked closely with engineering and operations to systematize the real world workflows into the operations interface. This complex decision matrix would become the backbone through which our entire team would operate off of.
The result was a highly customized and curated experience where the right task was presented to the right person, fully in context with every tool at the disposal of the user.
Outcomes and Next Steps
In order to measure the effectiveness of the taskboard and various features, we monitored real world working times for various tasks, both before and after implementation. In many cases, a 5-10 minute task could be completed instantaneously (such as in the case of the scheduling assistant). Over time, this would save operations several hours a day per person, amounting towards huge efficiency gains and cost savings. We measured an overall improvement of 30% time on task reduction.
Furthermore, these tasks aided by the additional info provided increased the accuracy and reduced support calls. Furthermore, new employees accelerated their onboarding routine and were able to effectively work on teams faster than before.
Lastly, one of the business needs was to open a second operations office. To reduce the burden of training, work management, and onboarding, this project was a critical requirement to the successful opening of that office.