Why parts data needs cleaning before a stock report means anything
A consumption report is only as good as the records behind it. In a workshop system the same part is often recorded several ways, and the report quietly counts each version separately. Before asking what to reorder, the data has to agree with itself.
How the same part ends up as five
- Brand names typed differently, sometimes in Arabic and sometimes in English
- Typing errors in part numbers and model references
- A genuine part and its aftermarket equivalents stored as unrelated items
- Part numbers replaced by the manufacturer over time
- The same export imported twice
What that does to a report
Demand is split across duplicates, so each line looks slow-moving. Stock of an alternative is ignored, so the report recommends buying a part you effectively already have. The totals look precise and are wrong.
Fix the labels first
Start by standardising brand and model labels. This is mostly rule-based work: a list of known spellings in each language, mapped to one name. Corrections to part numbers need more care and should rest on evidence, such as a matching manufacturer reference.
Group parts carefully
Parts that do the same job can be linked through trusted shared references. A catalogue match alone is not enough: two parts can share a reference and still differ in function. Grouping rules should check what the part does and leave uncertain matches for a person to decide.
Then calculate
With clean, grouped data, twelve months of invoice history gives a fair picture of demand. From there you can classify stock by value and by how steady the demand is, set stock cover and lead time, and produce reorder suggestions that count the eligible alternatives already on the shelf.
Keep people in the loop
Automation should propose, not impose. Every correction should be open to approval or reversal, and every buying suggestion should come with its reason. That is the approach we took for Handlers Auto, where preparation that took weeks by hand became a workflow the team runs on demand.