How many suppliers before geographic concentration risk actually drops

Procurement teams ask this one a lot during annual category reviews: we've got twelve suppliers now instead of four, so why does the risk register still flag the category as concentrated? Supplier count was never the number that mattered. Geographic spread was.
There's no magic threshold like "seven suppliers and you're safe." A category with forty vendors can still sit entirely inside a thirty-kilometer radius if they're all drawing from the same industrial estate or the same river delta's component makers. A category with five can be genuinely diversified if those five sit on three different continents, behind three different ports, in three different seismic zones. Count tells you how many invoices you'll process. It doesn't tell you what happens when one flood closes one road.
Why supplier count is the wrong metric
Procurement teams that built their diversification programs around headcount targets usually did it because headcount is easy to track in a spreadsheet. Lat/long isn't, at least not without a lot of manual cross-referencing against a map nobody has time to build by hand. So "add two more suppliers this fiscal year" became the proxy for "reduce our exposure," and in a lot of categories that proxy quietly broke down.
Here's the mechanism: when a region builds strength in a category, the supply base around it densifies. Tier-2 and tier-3 vendors cluster near the tier-1 anchor because of logistics, skilled labor, and shared infrastructure. A procurement team sourcing a connector or a casting or a specialty chemical can add a fourth, fifth, and sixth supplier to its panel and still be buying from within the same industrial park, because that's where the qualified capacity sits. The panel grows. The single point of failure doesn't move.
This is the gap between a vendor master and a concentration map. The vendor list says "diversified across twelve suppliers in six countries." The map says four of those twelve are two kilometers apart, downstream of the same river crossing, and the six countries include three with ports that all clear through the same strait.
What moves the number
A new supplier reduces concentration risk when it changes the geography, not the headcount. Three tests are worth running before you credit an addition toward your diversification target:
Does it sit outside the existing cluster's catchment, not just outside the existing cluster's postal code? A supplier forty minutes from your other three is still exposed to the same regional grid outage, the same port, often the same flood plain.
Does it depend on different upstream infrastructure? Two suppliers on opposite sides of a city can both be single-sourced from the same raw-material plant or the same power substation. The risk didn't move, it just changed its badge.
Does it clear a different chokepoint to reach you? If every supplier's freight transits the same strait, canal, or border crossing, your supplier count climbed and your exposure to that one chokepoint didn't.
If an addition fails all three, it's adding procurement overhead without buying down risk. That's a fine thing to know before the next sourcing round, and it's usually invisible until someone plots the panel against physical geography instead of a country field in the ERP.
The honest version of "minimum supplier count for diversification" is: there isn't one, independent of where those suppliers sit. A category with the right four is safer than a category with the wrong fourteen. What's worth tracking is how much of your spend sits inside any one radius, port catchment, or industrial cluster, and whether your last three sourcing decisions shrank that radius or just added names to the same patch of ground.
That's the cross-reference most category teams end up doing by hand once a year, pulling up a map and checking addresses against it one by one. Sourcing Concentration was built to do that pass automatically, plotting a category's spend against supplier geography and ranking the clusters it finds so the single point of failure shows up on a plan rather than staying buried in a spreadsheet of addresses.
If your next sourcing review needs to answer "did this diversify us" with something better than a gut check, that's the question worth running the map against.