Finding out about a shortage only when an order can't be filled?
Real-time visibility for a distribution network
Stock and shipments across every warehouse, seen as they change, not at the end of the week.
Network overview
This week, all regions
On-time delivery
94.1%
+1.3 ptsTarget 95%
Orders today
1,284
+6.2%Target 1,400
Vehicles on route
112
+8of 140 in the fleet
Stock cover
18.4 days
−2.1 daysTarget 21 days
Orders
Per day, this week against last
Live routes
112 vehicles across 5 hubs
Kigali97%
Nairobi96%
Johannesburg95%
Accra93%
Lagos89%
Deliveries by region
Completed this week
Activity
- Route LG-14 delivered 6 min earlyLagosJust now
- Stock below reorder point, SKU 1182Lagos depot2 min ago
- Route NB-07 delivered on timeNairobi5 min ago
- Delivery 3 h behind route planNairobi9 min ago
- Type
- Solution
- Sector
- Logistics & supply chains
- Capabilities
- Data & Analytics, Cloud & Platform
- Built for
- Distributors, wholesalers and logistics operators
§1Context and challenge
Context and challenge
A distributor runs several warehouses, each with its own stock system, plus an ERP for orders and a fleet tracked by a separate provider. Stock positions are compiled weekly from exports, and shortages are discovered when an order can't be filled.
The challenge
Give operations a live, trustworthy picture of stock and shipments across the network, and warn them about shortages before they happen, without replacing the operational systems that already work.
What changes
Decisions made on current data: fewer stock-outs and emergency transfers, and planning based on demand that is forecast rather than remembered.
What we deliver
Data architecture, change data capture and pipelines, data modelling, operations dashboards and demand forecasting.
§2Solution and architecture
Our solution.
Change data capture streams every stock movement and order change from the existing databases into a lakehouse, alongside fleet positions from the tracking provider's API. A modelled layer defines stock, demand and in-transit consistently across sites; operational dashboards and exception alerts sit on top, and a forecasting model projects demand per product and site.
Data platform
Data platform. Components:
- ERP (source): Orders
- Warehouse systems (source): Stock movements
- Fleet tracking (source): Provider API
- CDC + ingestion (process): Streams
- Lakehouse (store): Raw → modelled
- Operations dashboard (output)
- Exception alerts (output)
- Demand forecast (output)
Connections:
- ERP to CDC + ingestion
- Warehouse systems to CDC + ingestion
- Fleet tracking to CDC + ingestion
- CDC + ingestion to Lakehouse
- Lakehouse to Operations dashboard
- Lakehouse to Exception alerts
- Lakehouse to Demand forecast
§2.1The interface
What your team works in.
The interface we deliver, shown here with sample data.
Network overview
This week, all regions
On-time delivery
94.1%
+1.3 ptsTarget 95%
Orders today
1,284
+6.2%Target 1,400
Vehicles on route
112
+8of 140 in the fleet
Stock cover
18.4 days
−2.1 daysTarget 21 days
Orders
Per day, this week against last
Live routes
112 vehicles across 5 hubs
Kigali97%
Nairobi96%
Johannesburg95%
Accra93%
Lagos89%
Deliveries by region
Completed this week
Activity
- Route LG-14 delivered 6 min earlyLagosJust now
- Stock below reorder point, SKU 1182Lagos depot2 min ago
- Route NB-07 delivered on timeNairobi5 min ago
- Delivery 3 h behind route planNairobi9 min ago
Scroll sideways to see the whole interface.
§3Engineering decisions
The decisions that matter.
D1
Capture changes, don't replace systems
CDC reads each database's change log, so the warehouses keep their systems and the platform stays current within minutes.
D2
Define stock once
On-hand, reserved, in-transit and available are defined in one modelled layer that every dashboard and model uses.
D3
Alerts before dashboards
The first release ships exception alerts for projected shortages; dashboards follow. People act on alerts; they browse dashboards.
D4
Forecasts with ranges
Demand forecasts show uncertainty bands, and their accuracy is tracked, so planners know when to trust them.
§4Implementation approach
Implementation approach
01
Map
Inventory sources and agree definitions of stock, demand and in-transit with operations.
02
Stream
Set up CDC and ingestion for the ERP and the first warehouse; validate against existing reports.
03
Act
Launch shortage alerts and the operations dashboard; add the remaining sites.
04
Forecast
Introduce demand forecasting once enough clean history is in the platform.
§5How we measure success
The measures we agree with you before we start, and report against.
- Data freshness from source change to dashboard
- Stock-outs and emergency transfers per month
- Agreement between platform stock and physical counts
- Forecast error by product group
Have this problem?
Tell us how it shows up in your organization. We'll show you how the solution fits your systems and what it takes to put it to work.