Inside My Unilever Supply Chain Internship: Using KPIs and Root-Cause Analysis to Improve Inventory

During my supply chain internship at Unilever in Stockholm, I worked on turning inventory issues from vague signals into something trackable, discussable, and actionable. Rather than looking at one number in isolation, I focused on connecting stockouts, delivery performance, overstock, and replenishment decisions within one analytical framework so the team could work from a shared understanding.

Why Start with a KPI System

One common mistake in inventory work is treating outcomes as causes. When stockouts appear, teams immediately discuss replenishment. When overstock rises, the instinct is to cut inventory. But without consistent definitions, it is difficult to tell whether the issue comes from demand volatility, parameter settings, or execution gaps.

That is why I first helped build a clearer supply chain KPI system around a few core measures:

• Stockout rate: identify which SKUs are most exposed during critical periods.
• On-time delivery: track whether supply execution is stable instead of looking at inventory alone.
• Overstock exposure: locate slow-moving inventory that ties up resources.
• Weekly tracking cadence: keep all metrics in one rhythm so changes can be compared and discussed consistently.

The value was not just “one more dashboard.” It created a common language for problem solving. Once definitions are aligned, root-cause analysis and action prioritization become much more practical.

Breaking Shortages and Overstock into a Root-Cause Chain

Outcome metrics alone do not tell teams what to do next, so I built the analysis around a root-cause chain for both shortages and overstock. My approach was to start from the result, trace backward through the process, and identify which drivers were worth adjusting first.

I mainly focused on three groups of questions:

• Was there forecast error on the demand side, creating a mismatch between replenishment and actual consumption?
• Were replenishment settings such as reorder points or safety stock poorly calibrated?
• Were there execution issues such as unstable delivery, coordination delays, or misaligned operating rhythms?

This made discussions much more concrete. Instead of debating whether inventory was simply “too high” or “too low,” the team could distinguish whether the real issue came from forecasting, parameters, or execution, and then focus effort where it mattered most.

Turning Metrics into Actions

Once the KPI structure and attribution logic were in place, the next step was to convert analysis into specific actions. I focused on replenishment priority lists and parameter adjustments so weekly reviews could move beyond diagnosis and lead to decisions.

The key was not building a complicated model. It was helping cross-functional stakeholders align on the same numbers and interpret them the same way. Around replenishment parameters and safety stock, we gradually clarified which SKUs required stronger protection and which ones should carry less inventory, instead of applying one blanket policy to everything.

The outcome was not dramatic, but it was meaningful: as weekly tracking and parameter adjustments became more stable, overstock was reduced by 10%. To me, that result reinforced that inventory optimization is not just an analytical exercise. It is a coordination process that connects metrics, communication, and execution.

What I Took Away

This internship gave me a more practical view of supply chain analysis. Good analysis is not about making dashboards more complicated. It is about making numbers useful for decisions.

My main takeaways were:

• A KPI system is valuable when it clarifies problems, not when it adds more visuals.
• Inventory issues are usually system problems, so demand, parameters, and execution need to be viewed together.
• Effective analysis must end in action priorities rather than stopping at conclusions.

Looking back, the experience reshaped how I think about data work. Data does not improve operations on its own. What matters is turning data into a decision framework that teams can actually use, and that was the most important lesson I learned from my time at Unilever.