By Alma Yzaguirre, Safety Program Manager, Amazon
Operations managers are expected to keep safety top of mind in our warehouses, and I’m responsible for providing mechanisms and programs to make that happen. One way we strive to accomplish this is by getting out onto the operations floor at least once per month to observe conditions and behaviors, which may present a concern for employee safety, called “Safety Walks.” Think: a tripping hazard that could’ve been ignored for days or weeks (condition), or employees not wearing the required personal protective equipment (behavior). These are critical checks conducted in our org and the process includes:
- operations manager identifies the facility to conduct their walk
- manager conducts the safety walk and documents findings
- findings are submitted to a central repository for safety professionals to review
- safety professional identifies trends from the manager’s findings for which action plans are created to mitigate hazards
- manager “gets credit” for completing the safety walk
The expectation to complete safety walks was clear—they’re required and tracked on a weekly recurring basis, but the available tools weren’t capable of delivering the results we wanted. We were using a ticketing system which assigned the safety walk task to each manager globally. This required at least one hour for initial setup and had to be monitored to make sure the right people were getting assigned to the correct facility to complete the task. A global checklist that set expectations was established, but since the ticketing system didn’t enable embedding the checklist as part of completing the assigned task, we were left to hope that managers would access the checklist through a link provided in the assignment. Because managers weren’t forced to answer the standardized questions, the manager safety walk was applied differently by location. Some always followed the checklist, some interpreted the questions differently, some managers created their own checklists, too. These inconsistencies created additional challenges where the length of time to complete the safety walk varied by location, some took 10 minutes to complete while others took upwards of one hour or more. As every manager did their safety walks a little differently, we were unable to compare completed walks to each other without spending hours reading through each completed safety walk. This was time we didn’t have, and we couldn’t easily identify where to focus our attention to make the biggest positive impact to operations.
Making an impact, one step at a time
Because of the lack of standardization, the varied amount of time spent completing a safety walk, and the lack of measurable outputs, the safety walk process was not trusted. We needed a way to eliminate confusion around manager expectations, reduce the amount of time required to complete a safety walk, while also delivering meaningful outputs for the organization to take action against. And then we found Honeycode, which helped us achieve:
- standardized 11 questions sets, specific to work processes occurring on the operations floor for managers to select and complete
- reduced the time to complete a safety walk by 44% (down 20-minutes from 45-minutes, on average)
- identified site-specific and global improvement opportunities based on the trending data available to us; Personal Protective Equipment (PPE) not being used is one example
- the application framework enabled us to create additional workstreams related to pandemic response by creating a separate but equal COVID-19 inspection application
- provided transparency with operations with regard to at-risk behaviors and trends through the performance dashboard
- completed 2,519 safety walks (and counting!) in one geo-region alone, which is a 101% increase from the expected number of safety walks (1,253)
The app has been well-received across the org: 86% of users think using our app is easy, 83% prefer to use our app over the old process in the ticketing system. We extensively used the live update features, so as managers submit observations, the dashboard app is live updated. It gives us access to the most up to date information without having to go to multiple places. It also allows our app to be dynamic, so that a manager knows exactly what questions he or she needs to answer for a particular category before he or she even starts the observation.
Quote from Ops manager: “I think the best part about this tool is that it encourages Safety conversations with the team. Through my observations and the questions I was able to have many safety discussions on different topics with our teams. Thanks!” Shawn Boisvert, Logistics Manager
Quote from Safety Manager: “The manager observation tool provides us the data we need to focus on what is important as a DC H&S professional: what will make the work in the data centers safer. Through the observations we can work with Ops leaders to improve safety culture by rewarding safe behaviors, correcting unsafe behaviors, and learning what programs and training the employees need.” Keith Green, Regional Safety Manager
A glance at our Manager Safety Observations app
The dashboard app uses the summary tables views and app-user filterable detail screens extensively. This allows us to both look at high-level insights and dive into details when we see something is amiss. We also pushed the boundaries of what’s possible in Honeycode, for example we built our own calendar view using visibility conditions and lists provided in Honeycode. This gives us visual indicators as to whether managers are engaging in safety walks, and whether that is occurring every day.
Leaderboard - we generated a little friendly competition among each cluster to drive more completed observations. Users can toggle between the default All Time view and the current week view. Managers can also quickly access any incomplete observations if they get pulled into other work mid-walk.
Monthly Performance Report - safety professionals create a weekly report showing whether or not managers are completing the expected number of observations. Without knowing the power of Honeycode, they were looking for a place to extract the information from Honeycode into Excel format so they could create their reports. Instead, we built this monthly report to get a quick view of the managers' engagement. This also provided coaching and 1:1 opportunities to better understand why some managers were engaging in the new safety walk process (100% completion rate) versus others not engaging as much.
Summary by Region - we then realized other reports could be created, resulting in a bigger picture view of locations completion of observations as compared to the number of unsafe behaviors being identified. This started conversations around what some locations might be doing, above and beyond just safety walks, which could be introduced to other locations to reduce the number of unsafe behavior reports.
Multiple filtered view options - we also figured managers and safety professionals would want to look at every single question, if needed. We created multiple pages with filters enabled to parse the information in different ways. We can then evaluate responses by the category of question, by individual questions, by a facility, or by a manager completing the observation.
Our Honeycode journey ahead
We’re not stopping here—we’ll be expanding our app to the global operations by the end of 2020. We’ve already duplicated the framework of the safety walk application to facilitate our pandemic response plan by standardizing site assessment questions to ensure safety protocols are adhered to, and are actioned against in a timely manner where they are not. We’re considering using Honeycode for other safety workstreams such as: incident management, hazard assessments, project planning, proactive reporting, and more!