A Simple Way to Track Every Supplier's Reliability Over Time

A different kind of drift than price or stock

Checked 26 September 2026: our piece on the two spreadsheets worth building before you list anything covers the supplier log most sellers build first - a running record of a supplier's price and stock status, checked weekly, built specifically to catch the kind of drift that is invisible from memory but obvious once written down. It is exactly the right tool for what it tracks. It was never built to catch a different, quieter kind of drift: a supplier whose price and stock are both still perfectly fine, but who has started taking longer to reply, dispatching later against their own stated handling time, or needing an increasingly firm nudge before anything actually moves. None of that shows up in a price column. It only shows up if something is actually recording it.

The four things worth actually tracking

Behavioural drift has a shape, and it tends to show up in four specific places rather than everywhere at once.

Response time - roughly how many hours pass between a genuine message sent and a genuine reply, not an automated acknowledgement. A supplier who once replied within a working day and now regularly takes two or three is telling you something, even if every individual reply still eventually arrives.

Dispatch accuracy - whether an order actually left within the supplier's own stated handling time, checked against a real dispatch confirmation or tracking update, not simply assumed because nothing was said to the contrary. Our piece on what "handling time" actually means to eBay vs to you covers why this specific gap matters - a supplier's stated time and their actual, delivered time are not automatically the same thing, and only checking closes that gap.

Defect and quality consistency - genuine issues (damaged on arrival, wrong item, missing parts) tracked as a rate against orders placed that period, not a raw count. A rate stays comparable as volume grows; a raw count of "three problems" means something completely different at ten orders a month than it does at two hundred.

Communication pattern - a single, honest line each month on whether the supplier's general manner has shifted: still proactively flagging their own delays, or has that stopped happening, with delays only surfacing once chased? This is the most subjective of the four, and also often the earliest warning sign - a supplier who used to volunteer bad news and has quietly stopped is usually the same supplier whose numbers start slipping a month or two later.

What the actual log looks like

Nothing elaborate is needed here, and elaborate is usually what causes this kind of log to be abandoned within a month. One row per supplier, updated monthly rather than weekly - behavioural drift genuinely needs a longer window than price drift to mean anything; checking weekly mostly produces noise, since one slow reply in a fortnight tells you nothing, where the same pattern held across a month does. A simple green/amber/red score against each of the four categories, plus one short, specific line of note - "took three days to confirm a delayed dispatch, previously same-day" - is enough. The value is entirely in the discipline of updating it on a fixed date each month, not in the sophistication of the columns.

Why scoring from memory alone tends to flatter a supplier

A specific trap is worth naming directly: filling in a monthly score purely from memory, at the end of the month, tends to produce a more forgiving picture than the month actually deserved. Memory is disproportionately shaped by whichever interaction happened most recently or felt most dramatic, not by the honest average across the whole month - a supplier who had two genuinely slow weeks followed by one good one is easy to remember, at review time, as "fine really," simply because the most recent impression was a positive one. The practical fix is noting specific incidents briefly as they happen - a single line jotted down the day a reply takes three days instead of one - rather than reconstructing the whole month's pattern from an end-of-month impression alone. The monthly score is still the thing that gets reviewed and compared; the notes underneath it are what keep it honest rather than optimistic.

Why the trend matters more than any single month

A single amber score, in any one category, in any one month, is not a reason to act - suppliers have genuinely bad weeks for ordinary reasons, the same way a shop does. The pattern worth acting on is a category sliding from green to amber to amber-leaning-red across three or four consecutive months, the same logic our two-spreadsheets piece applies to a single stale price entry: one data point is not a log, and a log's entire value comes from being able to look back across several checks and see a genuine trend, rather than reacting to whichever month happens to look worst in isolation.

Applying this with one supplier, and with several

The log works slightly differently depending on how many suppliers a shop actually has. Our piece on the honest case for starting with just one supplier covers why a single, well-understood relationship is often the right starting point - and with only one supplier, this log's real value is mostly diagnostic: an early, honest warning that something is changing, well before it becomes a visible problem a buyer notices. With two or three suppliers in play, the log's role expands into a genuine comparison tool - laid out side by side, a monthly review across several suppliers at once takes minutes rather than requiring a separate, fresh judgment of each relationship from scratch, exactly the practical problem our piece on running three suppliers without any one slipping describes.

When this actually changes a decision

This log earns its keep at exactly the moments covered elsewhere on this blog. It is the evidence worth having before asking a supplier for a better deal - a genuine, tracked record of consistent reliability is a far stronger position to negotiate from than a general impression. It is very often the pattern that was quietly visible for months before a supplier eventually goes quiet altogether - not always, but often enough to be worth checking retrospectively when it happens. And it is the specific tool that makes running several suppliers without any one of them slipping actually practical, since a monthly score across four categories takes minutes to update and compare, where trying to hold the same judgment in memory across three separate relationships does not scale nearly as well. A supplier contract's stated handling time, covered in what a supplier contract actually needs to say, is also only as useful as the checking that holds a supplier to it - a number in a contract nobody is tracking against is a number that can drift quietly for months unnoticed.

What to actually do once a supplier scores red across several categories

A supplier who has genuinely slid to red, or amber-leaning-red, across two or more categories for several consecutive months warrants a specific, concrete next step, not simply another month of watching. The useful first move is a direct, factual conversation - not an accusatory one - grounded in the tracked record itself: "dispatch has been taking eight to ten days against your stated three for the last three months, and I wanted to check whether something has changed" tends to land better, and get a more useful answer, than a vague complaint about service getting worse. If nothing improves within a further month or two of that conversation, the honest next step is gradually easing order volume toward a second, better-tracked supplier - gradually rather than abruptly, so the shift itself does not create its own supply disruption - rather than either an abrupt cut-off or an indefinite continuation of the same slipping pattern.

Two worked examples

A supplier drifting the wrong way. A supplier scored green across all four categories in January: same-day replies, dispatch consistently within the stated 24 hours, a defect rate under 2%, and delays volunteered before ever being asked about. By April, response time has slipped to amber, dispatch has slipped to amber, the defect rate is still comfortably green, and communication has slipped to amber - delays are now only surfacing once chased. No single month's entry would have justified acting on its own. Looked at as a four-month trend, the pattern is unmistakable, and it is a genuinely useful moment to have a direct conversation, or start easing order volume gradually toward a second supplier, well before any of it turns into an actual failure a buyer notices.

A supplier genuinely improving. A newer supplier scored amber across response time and dispatch accuracy in its first two months - understandable while a new relationship is still settling in - before both categories move to green by month three, alongside a defect rate that has stayed green throughout. The same log that catches decline is just as useful for confirming a real, earned improvement, rather than a seller continuing to treat a newer supplier with the same early caution indefinitely, past the point the tracked record has actually justified more confidence.

Questions people ask