Logistic operations

A Network That Moved Goods, Not Value

At an operational level, the organisation faced persistent inefficiencies:

• stock imbalances across stores and warehouses

• frequent out-of-stock events in high-demand SKUs

• excess inventory in low-rotation products

• suboptimal routing and underutilised transport capacity

Despite having forecasting tools and ERP systems, decisions were made in silos:

• demand planning disconnected from logistics

• warehouse operations disconnected from transport

• pricing and promotions disconnected from inventory

The result was not failure.

It was systematic underperformance.

 

The Cost of Fragmentation

The organisation could not answer, in real time:

• Where is inventory actually needed right now?

• Which products should be prioritised in distribution?

• How should transport routes adapt to demand shifts?

• What is the real cost-to-serve per SKU and per location?

As a consequence:

• replenishment cycles were reactive

• safety stock levels were inflated

• transport costs increased unnecessarily

• working capital was tied up in the wrong places

 

Reframing the Problem

The issue was not forecasting accuracy alone.

It was the absence of a system capable of connecting:

• demand signals (store-level, real time)

• inventory positions (warehouse + store)

• logistics constraints (routes, capacity, lead times)

• commercial priorities (margin, promotions, turnover)

What was needed was not better reporting.

It was continuous optimisation of flow.

 

Building a Retail Logistics Twin

Cleohpatra implemented a real-time operational twin of the distribution network.

The platform integrated:

• POS data (store-level demand signals)

• warehouse inventory and throughput

• transport routes and fleet capacity

• supplier lead times and delivery variability

• SKU-level profitability and turnover

This created a unified model where every unit of inventory could be tracked, prioritised, and reallocated dynamically.

 

From Replenishment to Allocation

The most significant shift was in how inventory was managed.

Previously:

• replenishment followed static rules and forecasts

After implementation:

• inventory allocation became dynamic and value-driven

The system continuously:

• prioritised high-demand, high-margin SKUs

• rebalanced stock across locations

• adjusted replenishment based on real-time demand

• optimised load consolidation for transport

Logistics stopped being reactive.

It became decision-driven.

 

Optimising the Last Mile

Transport was another major source of inefficiency.

Before:

• routes were predefined

• trucks operated below optimal capacity

• last-minute adjustments were manual

After:

• routes were dynamically optimised

• load factors improved

• delivery windows adapted to real demand

This reduced both cost and delivery time variability.

 

The Shift in Performance

Within the first 6 months, the impact was measurable:

• +18% improvement in on-shelf availability

• -22% reduction in stockouts on high-demand SKUs

• -15% reduction in excess inventory

• -12% reduction in logistics costs (transport + warehousing)

• +9% improvement in inventory turnover

Most notably:

→ Working capital tied in inventory reduced by ~€85M

 

Speed as a Competitive Advantage

Equally important was the change in decision velocity.

• planning cycles reduced from weekly to near real-time

• cross-functional decisions (demand + supply + logistics) aligned instantly

• manual intervention reduced significantly

The organisation moved from:

• planning logistics

To:

• operating a live system

 

Beyond Retail

The case illustrates a broader principle.

In distributed systems, value is not created by moving goods.

It is created by moving the right goods, to the right place, at the right time — continuously.

 

What we think…

Retail logistics is often seen as an execution problem.

But at scale, it becomes a problem of intelligence.

The difference between cost and advantage
lies in whether the network is managed as infrastructure,
or as a system that thinks.