Košík Order Completeness

Context

Košík is a Czech e-grocery operator. A large share of assortment runs through Metro stores and warehouses. Order completeness - right items, right temperature, no damage - is how customers judge whether online grocery is worth trusting.

Problem

For an extended period, stock completeness and quality issues in the Metro collaboration blocked expansion into more cities and countries. Roughly 1 in 5 orders had some problem: missing items, wrong substitutions, broken eggs, warm frozen goods. The pain sat at the intersection of supplier, warehouse, systems, and customer experience - not in a single screen.

My role

I was accountable for fixing the problem end to end: diagnosis across countries, alignment of operations and product, and measurable improvement customers would feel.

Constraints

• Multi-country operations: Processes and ownership differed by site; fixes in one place did not automatically transfer.
• System vs floor reality: Availability in the app often disagreed with what pickers could actually find.
• No single owner: Warehouse, supplier, IT, and product each saw part of the problem.
• Expansion deadline: International rollout depended on reliability, not another pilot.

What I did

• Mapped data flows, floor communication, and when items became "available" in software vs physically stocked.
• Ran on-site work across markets; brought the same people into shared problem-solving who had never worked in one room.
• Drove process changes and product/system fixes where the handoff between digital availability and physical stock had no owner.
• Tracked problem-order rate as the north-star metric with operations and leadership.

Key decisions

• Process + product, not a hero feature: The fix was coordination and operational truth, not one UI tweak.
• On-site first: You cannot redesign e-grocery logistics from slides alone.
• Metric-led: Anchor on problem orders customers experience, not internal vanity counts.

Result

After months of cross-team work, problem orders improved from roughly 1 in 5 to about 1 in 100 - about a 20x reduction. Metro collaboration became viable for further expansion. Families ordering online were far more likely to receive what they paid for, in good condition.

What this shows

This project is a good example of the kind of work I do best: taking a messy operational reality (warehouse, supplier, and system gaps) and turning it into a reliable product and process system customers can actually trust.