Every manager scheduled shifts with no data.
Sede Café is one of the fastest-growing coffee chains in Mexico. But it had a problem that grew with every new store: each week, every manager decided how many baristas to put on each shift by intuition, with no data on their store’s real demand.
In July 2026 the imbalance was obvious. Half the stores had people to spare, paying for hours the demand didn’t ask for. The other half were short-staffed right at peak hours, which put sales and service at risk.
- Overstaffed stores
- Understaffed stores
- Pilot stores
Each store gets the staffing plan its demand calls for.
To solve this, we built a model that forecasts how many cups each store will sell, hour by hour, and combines it with its team’s productivity. With that and Sede Café’s operating rules, an optimization program calculates the lowest-cost staffing plan that covers demand: how many people, in which role and when they start.
Where there were people to spare, it asks for fewer; where the store was short, it asks for more. And the plan is recalculated every week.
Lower cost, same sales, better reviews.
To measure the impact, we compared stores with very similar characteristics: some ran on the model’s staffing plan, and the others kept operating as usual. Against their twins, the pilot stores achieved:
- −37% in operating cost
- +2.3 pts in 5-star reviews
- Steady sales, with no significant change
In short: they spend less on staff, sell the same, and their customers rate them higher.