How to spot a customer going quiet
Last updated 3 min readBy Allodial Predict
A customer rarely tells you they are leaving. What changes first is how they order from you: an order that does not arrive when it usually would, orders that keep coming but get smaller, or one product line that stops while the rest carry on. All three are visible in your own order history, account by account, if you measure each account against itself rather than against an average.
Why the usual reports miss it
A sales report sorted by revenue shows who bought the most. It does not show who stopped. An account that used to order every two weeks and has not ordered in six still looks fine on a year-to-date total, because its earlier orders are still in the sum. By the time the total looks wrong, the account has often been gone for a quarter.
The fix is to compare each account with its own past, not with other accounts. A quarterly buyer that has not ordered in six weeks is on schedule. A weekly buyer that has not ordered in six weeks is not.
Signal 1: the order that does not arrive
Every repeat account has a rhythm: the usual gap between its orders. When the gap since the last order runs well past that rhythm, the account is past the point where it would normally have reordered. That is a fact about its orders with you. It says nothing about why, and nothing about what the customer has on hand.
Allodial measures each account's rhythm from its own order dates and names how far past it the account is with three levels: Watch, Slipping and Gone quiet. There is no score. The exact rule is on the methodology page.
Signal 2: orders that keep coming, but smaller
Some accounts never stop ordering; they just send less of it your way. Compare the last 6 months of revenue with the same 6 months a year earlier. Allodial calls an account shrinking when that is down 40 percent or more, and first sets aside a single unusually large order so one big job a year ago does not make a steady account look like it is fading.
Signal 3: one product line stops
An electrical contractor that still buys wire from you but stopped buying fittings is still ordering, so no total looks wrong. Allodial flags line drift when an account bought a line in 4 or more of the prior 12 months, has bought under 25 percent of its usual amount of that line in the last 90 days, and is still at 50 percent or more of its usual spend overall. See product lines and clusters.
Signal 4: several accounts leave the same line together
When 3 or more accounts drift off the same product line within 60 days of each other, that is a cluster. One account leaving a line is a conversation with that account; several leaving together is often worth a look at that line itself.
The account that is already gone
The simplest check needs no rhythm at all: an account that bought in 4 or more of its first 12 months and has bought nothing in the last 3 months is lost. The free check runs exactly that rule, plus the shrinking rule, on a sales report inside your browser.
What to do with the list
- Work the accounts with the most revenue at stake first: the account's trailing 12 months of revenue.
- Call before you email. Ask what changed; do not guess a reason or name one.
- Log what happened, so the account leaves the list and nobody calls it twice.
- Come back to the ones you could not reach with a dated follow-up.
Allodial's daily list does the sorting each morning, capped so it stays workable, and Replay runs the same rules over your full history on the first day, so you can see which accounts drifted before you change anything.
Questions
How long before an account counts as going quiet?
It depends on the account. A weekly buyer is overdue after a few weeks; a quarterly buyer is not. Allodial measures each account against its own rhythm, which is why it names levels rather than a fixed number of days.
Can order history tell me why a customer stopped?
No. It shows what changed in their orders with you: when, how much, and which lines. The reason is something to ask them.
Where this shows up in Allodial
- Daily list Each morning, the accounts that broke their pattern
- Product lines and clusters Catch the account that stopped buying one thing
- Replay See what the rules find in your own history