Mid-market wholesaler aligned purchasing around one risk number

A mid-market wholesaler supplying installers and trade counters ran SAP Business One with disciplined master data and four buyers who still argued from four private spreadsheets every week. Wholesale inventory management came down to whoever argued best. Once every one of the 1,800 SKUs carried the same product risk score, with its components visible, purchasing moved from five days of debate to same-day decisions and cut emergency inter-branch transfers by 18% in a quarter.

The situation

Four buyers covered 1,800 SKUs across three branches and 400 key accounts: fixings, tools, sealants, and electrical consumables. Service-level targets were clear, at 97% line fill on A-class items. Execution was not.

Each buyer maintained an Excel model built on their own exports, their own velocity window, and their own definition of safety stock. Two buyers weighted supplier reliability; two did not. When a branch ran short, the fix was an emergency transfer from another branch, at freight cost and with a second branch now exposed.

Leadership had no single view of aggregate stock risk entering the month. The weekly purchasing meeting spent most of its two hours reconciling whose numbers were right before any purchase order was discussed.

What the data showed, and didn't say

SAP Business One stored every receipt, issue, transfer, and price change for nine years. Its dashboards showed turnover buckets and stock value by warehouse. They were accurate and nobody disputed them.

None of them answered the question that mattered on Monday: if we can only process 50 purchase orders this week, which 50 SKUs matter most? Ranking needs one consistent rule for velocity, cover, and lead time applied to every line, and that rule lived in four different spreadsheets.

Connecting Flowra

The team started with a CSV export pilot: sales lines, purchase lines, products, and clients for the previous 24 months, uploaded once. Flowra mapped the columns during onboarding and produced the first risk report within the hour. That was enough to convince the buyers the score reflected reality.

In week three, IT provided read-only credentials to the MSSQL database behind SAP Business One. Flowra connected through its SQL connector, refreshed nightly, and reconstructed a stock trajectory per SKU and branch from net transaction flow. SAP stayed the system of record, and nothing was written back.

What Flowra surfaced

• A product risk score on every SKU: 0.50 × slow-mover probability, 0.30 × velocity rank, 0.20 × trend rank, with the weighted components shown next to the number.
• A separate stockout signal per SKU and branch: critical when days of supply were within one supplier lead time, reorder soon within two.
• A weekly reorder queue ranked by revenue at risk and capped to what receiving could process, with reorder quantities from average daily sales, lead time, and a 95% service level.
• Account-level flags where demand on historically stable lines had dropped, from the client RFM and spend-collapse signals.

What the team did

Management set operating thresholds with the buyers rather than for them: any SKU in the critical tier, or with a critical stockout signal, went to a daily 10-minute stand-up; high-tier lines went to the weekly buyer block; medium and low tiers were monitored through the Monday digest.

Buyers kept the right to override. In the first month they overrode 7% of recommendations, mostly where a supplier had an unannounced minimum order change. Each override was logged with a reason, and the logged reasons updated lead times and pack sizes in SAP within the month.

The four spreadsheets were retired in week five. Not because anyone banned them, but because the shared list already contained the evidence they had been built to produce.

Results

"The risk score gave our purchasing team a single number to align around. Decision-making went from weekly debates to daily actions."

— Supply Chain Manager, mid-market wholesaler

What made it work

• One score, components visible. The buyers accepted the number because they could see the slow-mover probability, velocity rank, and trend rank behind it. How the score is built is covered in stock risk score explained.
• Thresholds tied to cadence. Critical meant today, high meant this week, medium meant the digest. Nobody had to decide how urgent a number was. See stockout risk windows.
• Overrides as data. Logged exceptions fixed master data faster than any cleanup project. The meeting format is in evidence-backed inventory meetings, and the principle behind it, that Flowra prepares and the team decides, held on every line.