Case Study

Forecasting the Future

How a national manufacturer cut forecasting labor by 89% — and finally trusted their numbers.

Industry
Bulk Goods Manufacturing and Distribution
Engagement
Custom forecasting module built into an existing ERP system
Client
Anonymized — national producer and distributor of landscaping and bulk goods, 10+ facilities across the U.S. and Canada
Relationship
Ongoing — maintained for 9+ years
89%
Reduction in forecasting labor
~10%
Forecast error (down from 25–50%)

The Situation

Our client was one of the larger suppliers of bulk landscaping goods to major home improvement retailers across North America. With more than ten production facilities spread across the U.S. and Canada — each specializing in different product lines — and a business that ran at dramatically different volumes depending on the season, accurate demand forecasting wasn't optional. It was existential.

The problem: they couldn't do it. Not reliably.

The Challenge

Year after year, the company ran the same painful cycle. As peak season approached, regional teams would begin pulling their people — not to prepare for the season, but to build forecasts for it. Twelve people. Six weeks. Spreadsheets. Historical data of questionable accuracy. Manual reconciliation across regions, facilities, and product lines.

At the end of those six weeks, they'd have numbers. And those numbers would routinely be wrong by 25 to 50 percent.

The gap between projected and actual demand produced inventory crises in both directions — too much in some regions, not enough in others — and eroded profit margins at exactly the moment the business could least afford it. As the company grew, the cost of inaccuracy grew with it.

What We Built

Rock Agile designed and built a forecasting module integrated directly into the client's existing ERP system. Rather than requiring teams to build forecasts from external spreadsheets and manual reconciliation, the module pulled from live operational data already in the system — product lines, facility output, historical order patterns, regional demand signals — and generated forecast models for teams to review and adjust.

The goal wasn't to eliminate human judgment. It was to eliminate the six weeks of manual data assembly that preceded it.

The Results

The impact was measurable from the first forecasting cycle.

What had taken 12 people six weeks — across regions — now took 4 people two weeks. The labor reduction was 89%.

More importantly, the numbers became trustworthy. Estimated forecast error dropped from a range of 25–50% to approximately 10%. Regional teams could make inventory decisions with confidence rather than educated guesses. The downstream effects — on inventory levels, on labor planning, on logistics — followed.

The system has been in use, and maintained by Rock Agile, for over nine years.

The Long View

That last detail matters as much as the numbers. We didn't deliver a feature and move on. We've been maintaining this client's system through ownership changes, through seasonal peaks, through the long arc of a business growing and adapting. That's what Rock Agile is built to be: the team that's still there.

Working on something similar?

Tell us about your situation and we'll tell you honestly whether we can help.

Get in touch →