Most distributors find out an account is slipping when the rep notices, or when the year-end numbers come in. By then the customer has often moved part of their spend somewhere else. You can catch it much sooner with data you already have: your invoice lines.
The idea in one paragraph
Every regular customer has a rhythm for each item family they buy. Some reorder fasteners every two weeks and safety supplies every six. Work out each account’s normal gap between orders for each item family, then flag the ones that are well past their own normal. The rep gets a short weekly list with the reason attached.
What you need
- 12 to 24 months of invoice lines from your ERP: customer, ship-to if you track it, item, item family, invoice date, quantity
- A way to export that weekly, like a saved report
- The rep or CSR who owns the accounts, for about an hour a week while it’s being built
If customer or item IDs don’t match across systems, mapping them is part of the work.
The method
- Group the history by customer and item family. Item family matters. An account can keep buying one line while quietly moving another.
- Find the normal gap. For each customer and item family with enough orders, work out the typical number of days between orders. A median is steadier than an average because one odd order doesn’t throw it off.
- Measure today’s gap. Days since the last order for that customer and item family.
- Flag the slips. Flag it when today’s gap is well past the normal gap. Start with a simple multiple and tune it after a few weeks of rep feedback.
- Show the reason. Each flag shows item family, last order date, expected interval and days overdue, so the rep knows what to ask about.
- Close the loop. The rep logs what they heard: lost to a competitor, project ended, seasonal, still buying. That feedback makes next month’s list better.
Where it breaks
- Customers who order rarely don’t have a predictable rhythm. Label them “not enough history” instead of guessing.
- Your own stockouts can make it look like the customer stopped buying. Check fill rate before calling.
- Seasonal items need a longer look-back, or a comparison to the same period last year.
- Project buyers order in bursts. They belong on a different list.
What it should not do
It shouldn’t email the customer or change a price. It gives a person a short list and the reason. The person decides what to do.
Doing it yourself vs. having it built
A capable analyst can do a first version of this in a spreadsheet from an ERP export. Where it usually stalls is keeping it running every week, getting it in front of reps in a form they’ll actually use, and feeding their notes back in. That’s the part a Pilot Build covers: about 30 days, fixed fee, training and a runbook at the end.
More first-project ideas for distributors: AI for distributors. Or book a 30-minute call.
Want to see it working? Here’s a sample build on invented data: the weekly list, the reasons, and the email a rep would get.
