Case Study · Mars Snacking × AiOn
How a major CPG distributor used OneTrack's AiOn agent to surface $60K+ in labor cost exposure across a national network of distribution centers, and cut costs nearly 28% in the first two months of deployment.
Mars Snacking, a major CPG distributor operating a national network of distribution centers, runs OneTrack's AiOn agentic Weekly Lost Time Analysis to review indirect labor costs and break compliance every week. The agent surfaced $60K+ in labor-cost exposure and drove a nearly 28% cost reduction in the first two months of deployment.
Inside the AiOn Weekly Lost Time Report at Mars Snacking.
The Situation
For a major CPG distributor operating a national network of distribution centers, indirect labor costs (time spent on breaks, sanitation, pallet transfers, and leadership-directed tasks) were accumulating every single shift. But for most of the year, that cost was invisible.
Weekly reporting from third-party logistics partners provided directional context. But the true cost picture only emerged at period close, when finance surfaced the variance. By then, weeks of opportunity had already passed.
Supervisors who wanted to dig deeper faced a time problem. A thorough manual analysis of indirect labor and break compliance could take one to two hours per shift, per supervisor, or five to ten hours of analytical work per week. Work that produced insight too slowly to change behavior in time.
"We received weekly directional reporting from our 3PL providers, but the true cost came in at period close."
The Solution
Mars Snacking adopted an agentic Weekly Lost Time Analysis, an automatic weekly labor cost review of their distribution network. This detailed calculation of indirect labor hours and costs by work code, with break and lunch violations down to the individual employee, delivered the full breakdown to DC managers and 3PL partners on a regular cadence.
No one has to pull the data. No one has to build the spreadsheet. The analysis that used to take hours now takes zero, because the agent runs it automatically and delivers it directly.
Persistent intelligence, delivered weekly:
"The biggest takeaway is the time supervisors have gotten back. Typically this type of manual analysis can take hours, especially for larger shifts. Now this is delivered daily to the 3PLs, weekly to the DC managers, and everyone can spend more time taking action than analyzing."
The Results
The first report identified over $29K in labor cost exposure across the network, a figure that had previously been invisible until period close. Over the first two months of deployment, that number dropped by nearly 28%: a sustained reduction driven by direct action from supervisors and 3PL partners who now had real data to act on every week. The difference was specific, real-time data pointing to true areas of loss, in time to change behavior within the same period.
"3PL conversation and supervisor action helped lead to our drop and our continued focus. We talk productivity with our 3PL teams weekly, and now have real data pointing to true areas of loss."
The coaching impact extended beyond cost. Because the report surfaces exact instances of when and how individual operators are performing below goal, supervisors can now have specific, evidence-based coaching conversations. The result: an 8.5% increase in units per hour at one of the network's lower-performing distribution centers.
"We are seeing increases in units per hour because OneTrack now gives supervisors exact instances of when and how an operator is below goal for better coaching."
By the numbers
The Bigger Shift
The Weekly Lost Time Report changed more than the numbers. It changed the nature of the weekly 3PL conversation. Where those calls were previously driven by directional estimates and period-end reconciliations, they are now anchored in specific, real-time data. Both sides arrive with facts. The conversation moves faster and further.
And the appetite for more is growing. Leadership at Mars Snacking understands what's possible when operational data is connected, automated, and delivered in context.
"The benefit for more data sources is becoming more and more clear for leadership."
What This Means
Before the agent, a supervisor spent one to two hours per shift building the picture. After the agent, that time is zero. The picture is more complete, more timely, and more actionable than anything a manual process could produce.
The cost reduction is real. The time savings are real. The productivity improvement is real. But the most durable change is that Mars Snacking's operational leaders now have a system that sees everything, forgets nothing, and delivers the right information to the right person at the right time, every single week.
The AiOn agent surfaced more than $60K in labor-cost exposure across a national DC network and cut costs by nearly 28% in the first two months since go-live. The first report alone identified over $29K in labor-cost exposure that had previously been invisible until period close.
AiOn is OneTrack's persistent intelligence layer, and here it runs an agentic Weekly Lost Time Analysis: an automatic weekly labor-cost review of the distribution network. It calculates indirect labor hours and costs by work code, flags break and lunch violations down to the individual employee, surfaces post-break idle time, ranks top associates by lost time per location, and provides a two-month cost comparison, delivered daily to 3PLs and weekly to DC managers.
A thorough manual analysis of indirect labor and break compliance previously took one to two hours per shift, per supervisor, or five to ten hours of analytical work per week. With the agent, that manual analysis time is now zero because it runs automatically. In a single week the agent tracked, named, and costed 800+ break violations.
Yes. Because the report surfaces exactly when and how individual operators perform below goal, supervisors can hold specific, evidence-based coaching conversations. This produced an 8.5% increase in units per hour at one of the network's lower-performing distribution centers.