Dairy & Frozen Desserts FMCG
Schedule overlap rate achieved
Time to generate a full week of production plans
About
With summer peaks looming, the business needed a cockpit that could craft collision-free production schedules and stress-test any what-if scenario in minutes. Overlapping runs and freezer bottlenecks cost Rs. 46 lakh in changeovers each season, and 30-day raw-material lead times left planners flying blind. We built a control tower backed by a branch-and-price scheduler that honours blast-freezer curves.
Industry
Dairy & Frozen Desserts FMCG
Company size
1,000 – 5,000 employees
Founded
1907
The Company
A seasonal dairy business planning blind into its own peak
The business manufactures dairy and frozen dessert products with a sharply seasonal demand curve that peaks hard through the summer months. Production runs across multiple lines, constrained by blast-freezer capacity that behaves very differently from a standard packaging line.
Raw-material lead times of roughly 30 days meant planners had to commit to production plans well before they had clear visibility into how the peak would actually unfold.
The challenge
Freezer bottlenecks and blind commitments cost Rs. 46 lakh a season
Overlapping production runs and freezer capacity bottlenecks cost the business roughly Rs. 46 lakh in changeover inefficiency every peak season, as manually built schedules struggled to sequence runs around the blast-freezer's specific thermal constraints.
With 30-day raw-material lead times, planners were locking in commitments without a reliable way to stress-test how those commitments would hold up against different demand scenarios — a monsoon delay, a promotional spike, a supply disruption — before it was too late to adjust.
The Solution
A branch-and-price scheduler with a 200-scenario swarm on demand
We built a Tableau-based control tower backed by a branch-and-price scheduling engine that respects blast-freezer thermal curves as hard constraints rather than an afterthought, producing collision-free schedules directly. An AWS Batch swarm can spin up 200 scenarios on demand, logging every assumption for audit.
Planners use the control tower to explore what-if scenarios — a demand shock, a line going down, a raw-material delay — and see the schedule impact within minutes rather than committing blind and discovering the consequences weeks later.
The Results
A full week of plans in 90 seconds
Schedule overlap fell below the 2% target, and a full week of production plans now generates in 90 seconds. Planners can explore five full what-if scenarios in under five minutes — a task that previously would have meant days of manual rework to even approximate.
Going into the peak season with this level of scenario visibility changed how the planning team approached raw-material commitments, replacing gut-feel bets with tested, auditable plans.
KEEP READING
Explore more studies on AI, analytics, and enterprise intelligence.
