Purchasing and operations analyst

Supplier price lists by email into one ERP CSV format

Give each supplier its own pipeline inbox. Every XLS or XLSX attachment is converted with a saved mapping template that renames columns, cleans prices and dates, validates required fields and delivers one consistent CSV to the SFTP folder or bucket your ERP imports from.

Suppliers send price lists in their own layouts. One file has the header on row 6 below a logo, another calls the SKU column Art.-Nr., a third writes prices with a decimal comma and a currency symbol. Your ERP accepts exactly one CSV layout, with fixed columns, a fixed delimiter and ISO dates.

Today someone downloads each attachment, fixes it in Excel and uploads the result. It takes time every week, depends on one person knowing each supplier quirk, and a single shifted column can push wrong prices into purchase orders.

The workflow

1. 1Receive the fileEach supplier sends to a dedicated pipeline address, or your mail system forwards their messages there. XLS, XLSX, XLSM, ODS and CSV attachments are accepted.
2. 2Find the header rowThe template locates the header by a known column name, so extra rows added at the top of the file do not break the mapping.
3. 3Map and clean columnsSource columns are renamed to your ERP fields, unused columns are dropped, prices lose currency symbols and decimal commas, SKUs are padded and dates are written in ISO format.
4. 4ValidateRequired columns, non empty SKUs, positive prices and a plausible row count are checked. A failing file is held and the owner is notified with the reason.
5. 5DeliverThe finished CSV goes to the SFTP folder, S3 bucket or webhook your ERP import reads from, in UTF-8 with the delimiter it expects.

An example calculation

Example calculation, not a customer result. If an analyst spends 4 hours per week cleaning supplier files at a loaded cost of 45 USD per hour, that is 180 USD per week, or about 9,360 USD per year. The Business plan billed yearly costs 888 USD, so the time saved covers the subscription many times over, before counting errors avoided.