Most one-person DTC teams learn their supplier has a problem the day a PO ships late — by which point the reorder is already off the rails and the empty-shelf window has begun. The dashboard a founder actually has rarely says anything about supplier health: it shows the open POs, the inbound dates, and a vendor list that hasn’t been re-ranked since onboarding. It doesn’t surface the vendor whose fill rate quietly slid from 96 to 84 over a quarter, or the one whose SLA slipped twice in a row without anyone flagging the pattern. The signal is real; the surface for it isn’t.
Vendloom’s supplier-health loop watches three numbers at once — late-shipment rate per vendor, fill-rate drop on the most recent PO, and the time delta between the promised and confirmed date — and decides whether a chase is already warranted, or whether one is going to be warranted soon. The loop lives in the same place the supplier confirmation lands, so a missed SLA doesn’t have to travel through Slack, Notion, or a Friday-morning inbox check before a human sees it. The chase drafts itself, and the founder’s attention is reserved for the cases where two signals agree — not for every tiny wobble in the inbound calendar.
What supplier-health surfaces before you do
- Late-shipment rate climbing past 8% on a vendor that was steady at 3% — early warning before the next PO is six days late and the inventory plan falls apart.
- Fill-rate drop on the most recent PO, even on a vendor that’s otherwise reliable — a 96 to 84 slide over a quarter is the kind of slow-drift the eye skips but the books absorb.
- Supplier SLA flags, meaning a missed confirmed date on two of the last three POs — a single slip is noise; a pattern is a vendor problem worth raising.
- Repeat-vendor slippage on the SKUs already short on cover — a late vendor on a SKU you’re understocked on is the highest-priority version of the same signal.
What the loop does — and what it does not
On a healthy beat the loop drafts the chase, threads it into the existing supplier thread, and updates the PO log — the founder reads nothing. The ping lands only when two signals agree across the same vendor over the same window: a late-shipment climb paired with a fill-rate slide, or an SLA flag coupled with a thin cover on the affected SKU. The point isn’t to draft more chases. It’s that the chases you already owe are the ones that arrive on time — and that a vendor problem is the kind of thing the founder hears about once, not five times, when the loop earns the right to draft the work between pings.
“A vendor that ships on time for a year is invisible. A vendor whose fill rate quietly slides to 84 is the same year the inventory plan slides with it.”
If two signals agree on a vendor over a single quarter, the loop treats the chase as a real assignment and pings the founder with the message already drafted and the supplier already pinged. The cost of doing it the other way is a vendor problem you find on a Tuesday, three weeks after your reorder landed late and the SKU went empty. Together with the stockout-risk and pause-underperforming-ads loops, supplier-health is the third side of the same posture: a loop that drafts the boring work quietly, reserves the founder’s attention for the calls that genuinely need a person in front of them, and turns supplier follow-through into the same ‘fewer signals, real signals’ discipline the other two already enforce.