Routing Optimization for a Fortune 500 Company
Fixing the data foundation, then optimizing what runs on top of it.
Industry: Fortune 500 Retailer
Scale: National Real-Time Optimization
The Challenge
A Fortune 500 company needed to improve operational efficiency through routing. The existing routing solution provided a single fixed path through each location, regardless of which locations needed to be visited.
The opportunity was clear: optimized routing could dramatically reduce travel distance and time. But the real challenge wasn't just the algorithm. It was the data.
The Routing Problem
A one-size-fits-all approach meant that every route followed the same path, an inefficiency costing distance and time.
The Data Problem
Floorplan layouts existed for other purposes and had never been validated for routing. Outdated configurations, missing sections, and inconsistencies were a significant challenge.
The Process Problem
The teams that owned the layouts had never needed to maintain them for routing accuracy. Bringing them along required careful stakeholder management and new processes.
Step 1: Fix the Foundation
Before we could optimize anything, we had to solve the data quality problem. The floorplan layouts were originally built for planning and merchandising purposes, not for high-stakes operational routing. When we started using them as inputs to a routing algorithm, quality issues that had been invisible suddenly became critical.
What We Found
Layouts that had been “good enough” for their original purpose were full of problems when used for routing: outdated fixture configurations, missing sections, inconsistencies between the digital layout and the physical reality. And any one of them could produce an invalid or inefficient route.
What We Built
We built automated validation checks and reporting that flagged invalid or suspect configurations before they could impact production routing. This gave the layout teams clear, actionable feedback: here's what's wrong, here's where it is, here's what needs to change.
- Automated checks that validated layout configurations against routing requirements
- Reporting dashboards that surfaced issues to layout teams before they hit production
- New processes and quality standards for layout maintenance, built collaboratively with the teams that owned them
The Stakeholder Challenge
This was as much a process problem as a data problem. The layout teams had maintained their data for years without issues because the old routing system didn't need precision. Now we were asking them to meet a higher standard for a use case they'd never supported. Success required close collaboration: understanding their workflows, building tools that fit into their process, and demonstrating that the quality improvements would benefit everyone, not just the routing team.
Step 2: Optimize the Routing
With clean, validated layout data in place, we could tackle the routing problem itself. The existing solution was straightforward: a single predetermined path through the floorplan. Every route followed the same path regardless of which locations needed to be visited. For smaller routes, this meant walking unnecssary sections. Across thousands of locations and millions of routes, the small inefficiencies added up.
Before
One fixed path through the entire floorplan. Every route, regardless of size or locations, followed the same general route from start to finish.
After
Optimized routes tailored to each need, minimizing travel distance based on the specific locations the route needed to visit.
Understanding the Constraints
Routing optimization at this scale isn't just about finding the shortest path. We had to carefully understand the constraints of the problem: physical layout limitations, operational rules, equipment restrictions, and crucially, which operational changes were actually on the table.
The best theoretical solution doesn't matter if it requires changes the operation can't support. We spent significant time with operational teams understanding what was feasible, what was flexible, and what was non-negotiable. That understanding shaped the optimization approach and ultimately made the difference between a model that works on paper and one that works in practice.
The result was a solution that respected real-world constraints while still capturing the vast majority of the theoretical savings. Optimized routes reduced travel distance significantly across every location, translating directly into reduced operational costs.
The Results
Millions of Dollars in Annual Savings
Optimized routing reduced travel distance across thousands of locations. At scale, even small per-route improvements compound into massive savings. The primary driver was straightforward: less walking means less time, which means lower operational costs.
A Data Foundation That Didn't Exist Before
The automated validation and reporting we built didn't just enable this project. It created an ongoing quality assurance process for layout data that the organization didn't have before. Layout issues are now caught and fixed proactively, not discovered in production.
Operational Teams Brought Along
The layout teams went from maintaining data for passive planning purposes to owning a critical input for production routing. They have the tools, the processes, and the understanding to maintain quality on their own. The improvement is sustainable because the people closest to the data own it.
What Made This Work
This project could have been approached as a pure optimization problem. But the millions in savings only materialized because we treated it as a data and process problem first.
Data quality before optimization
The best algorithm in the world produces garbage if the inputs are wrong. We invested heavily in understanding and fixing the data foundation before writing a single line of optimization code.
Stakeholder management as a first-class concern
We didn't just build tools and hand them over. We worked alongside the layout teams to understand their workflow, build solutions that fit their process, and make sure the new quality standards felt achievable rather than imposed.
Understanding real-world constraints
We spent time with operational teams to understand which changes were feasible and which weren't. That shaped an optimization approach that captured maximum value while respecting the realities of how the operation runs.
Have a similar challenge?
Whether it's a data quality problem, an optimization opportunity, or both, we'd love to hear about it.
Get in Touch