Case studies

Real results from companies that chose to decide with data

Data integration, management dashboard and predictive modeling projects, told through the numbers that matter to each client.

  • Consumer goods

A single version of sales for the entire distributor network

Leading consumer goods company in the potato chip market

distributors integrated into a single model
40
The challenge
Each distributor sent its sales in spreadsheets with its own formats and criteria. Consolidating them took days of manual work and leadership didn’t fully trust the final number.
What we did
We automated how the files are received, unified product and customer codes, and published a performance dashboard with filters by region, channel and category.
The outcome
Leadership and the sales team work with the same number and spot early which distributor or channel is drifting away from target.

Sources integrated

  • Excel and spreadsheets
  • .txt and .csv files
  • ERP

Technology

  • Power BI
  • Azure
  • SQL
Sales performance · Distributor network Sample data
  • Period: Jan–Dec 2025
  • Region: All
  • Channel: All
Net revenue
$48.6M ▲ 9.2% vs. 2024
Volume sold
1,284 t ▲ 4.1% vs. 2024
Average order size
412 kg ▼ 2.3% vs. 2024
Active customers
6,940 ▲ 3.8% vs. 2024

Monthly revenue vs. target

$ millions

  • Actual
  • Target
View data
MonthActualTarget
Jan$3.6M$3.8M
Feb$3.4M$3.6M
Mar$3.9M$3.9M
Apr$3.8M$3.9M
May$4.0M$4.0M
Jun$4.1M$4.1M
Jul$4.3M$4.2M
Aug$4.2M$4.2M
Sep$4.0M$4.2M
Oct$4.3M$4.3M
Nov$4.4M$4.4M
Dec$4.6M$4.5M
Illustrative dashboard with sample data, similar to the one we built. See the interactive dashboard
  • Distribution and retail

Purchasing planned with a demand forecast

Wholesale distributor

fewer stockouts
12%
The challenge
Purchasing was planned with historical averages: some products ran out of stock at peak times while others piled up in the warehouse.
What we did
We trained a model that combines sales history with seasonality and promotions, and integrated it into a weekly replenishment dashboard.
The outcome
The purchasing team knows what to restock, and how much, before it runs out.

Sources integrated

  • ERP
  • APIs and e-commerce

Technology

  • Python
  • Azure
  • Power BI
Demand forecast · Central warehouse Sample data
  • Horizon: 8 weeks
  • Category: All
  • Warehouse: Central
Forecast accuracy
92.4% ▲ 6.1 pp vs. previous method
Forecast demand (8 wks)
231,600 units ▲ 12.5% vs. last 8 wks
Products at risk of stockout
37 ▼ 12 vs. last week
Estimated overstock
$2.3M ▼ 18% vs. last quarter

Weekly sales and forecast

Thousands of units

  • Actual
  • Forecast
  • Likely range
View data
WeekActual (thousands)Forecast (thousands)Likely range
Week 121.2——
Week 222.0——
Week 321.6——
Week 423.1——
Week 522.8——
Week 624.0——
Week 723.5——
Week 822.9——
Week 924.6——
Week 1025.1——
Week 1124.3——
Week 1225.8——
Week 1326.2——
Week 1425.5——
Week 1526.9——
Week 1627.4——
Week 17—27.927.1 to 28.7
Week 18—28.327.2 to 29.4
Week 19—27.626.3 to 28.9
Week 20—29.027.4 to 30.6
Week 21—29.427.6 to 31.2
Week 22—28.826.8 to 30.8
Week 23—30.127.9 to 32.3
Week 24—30.528.1 to 32.9
Illustrative dashboard with sample data, similar to the one we built. See the interactive dashboard

Want results like these in your company?

Tell us about your challenge and we'll map out the first step together. No cost, no commitment.

Book a call Message us on WhatsApp