Industrial supplier mapped five years of Odoo history in one morning

A specialised industrial supplier sold maintenance, repair, and safety SKUs to about 300 factories. Odoo had been the system of record for five years, yet every board meeting still began with finance rebuilding an ad hoc Odoo inventory analysis in spreadsheets. IT enabled read-only API access one afternoon. By the next morning Flowra had produced a product risk report across the full five years, and the CEO used it in that week's board pack.

The situation

The catalog held 5,800 SKUs: gloves, filters, abrasives, lubricants, and a long tail of spare parts kept for specific customer machines. Most lines were cheap. A few were critical: if a plant's filter cartridge was out of stock, a production line waited.

The CEO wanted an honest picture of catalog health before each quarterly board: dead stock, margin leaks, and lines that looked fine on the shelf but were trending the wrong way. Internal BI was backlogged. A consultancy had quoted a three-month data warehouse project to answer questions that changed every quarter.

Finance spent the better part of two weeks per quarter exporting Odoo reports, joining them in Excel, and defending assumptions in the meeting instead of discussing decisions.

What the data showed, and didn't say

Odoo contained quotations, sales orders, deliveries, purchase receipts, and invoice lines with cost snapshots at the time of sale. The standard reports were organised around accounts and periods: revenue by customer, stock value by category, margin by product for a chosen month.

They were account-centric rather than decision-centric. No single document ranked products by combined operational and financial risk, and no report explained why a product with stable list prices was earning less each quarter. The cost snapshots held the answer, but nobody had time to read 60 months of them.

Connecting Flowra

IT created a read-only Odoo user with an API key. Flowra connected to Odoo 17 over XML-RPC through the Odoo connector and ingested five years of sales, purchase, stock-move, and invoice-line history that afternoon.

The overnight pipeline reconstructed a stock trajectory per SKU from net transaction flow, computed the product risk score and stockout signal for every line, and read the margin rate from the invoice history. The report was waiting the next morning. Odoo was never written to, and nothing about its configuration changed.

What Flowra surfaced

• A full product risk report: every SKU scored 0.00 to 1.00 with its slow-mover probability, velocity rank, and trend rank shown, and 412 lines in the critical tier.
• The top 30 SKUs with negative margin drift despite stable list prices, from the margin-squeeze signal. In most cases the cause was purchase cost creeping up while a volume-tier discount stayed fixed.
• Reorder recommendations weighted by supplier lead-time variance from purchase-to-receipt history, so the lines with unreliable suppliers carried more safety stock and the stable ones less.
• A dormancy list: 880 SKUs with no movement for over 90 days, ranked by cash trapped, and a separate dead-stock onset watchlist for lines still selling but likely to go silent.

What the team did

The CEO took the risk report and the top-30 margin list into the board pack unchanged, with Flowra's explanations as the appendix. The board discussed three decisions instead of the numbers: a clear-down budget for the dormancy list, a repricing round on the 30 lines, and a supplier review for the two vendors driving the cost creep.

The operations lead exported the reorder recommendations into the weekly supplier call sheet. Finance stopped rebuilding the analysis and instead reviewed the Monday digest, asking Flowra in Slack for the evidence behind any line that looked surprising.

Sales challenged four of the 30 margin flags, where a rebate arrived through a separate credit note. The challenges were logged, the rebate lines were added to the ingestion, and the four flags cleared on the next nightly run. The score improved because a human corrected it.

Results

"We connected Odoo in one afternoon. By the next morning, we had a complete product risk report for the past five years."

— CEO, specialised industrial supplier

What made it work

• The ERP stayed the ERP. Odoo kept doing what it does well: transactions, valuation, audit trail. Flowra became the decision layer on top. The split is explained in ERP analytics versus ranked action lists.
• The board got decisions, not data. A ranked list with explanations turned a two-hour numbers review into three decisions. Each line carried the fact, the forecast, the recommendation, and the hypotheses.
• History did the work. Five years of invoice-line cost snapshots exposed margin drift that a monthly report could not. How the score reads that history is in stock risk score explained.