Supplier Concentration and What a Top Five List Hides

Why this matters

A top supplier list ranks your suppliers by what you committed to them over a window, and every owner already knows the order. You could name your top three without looking. What you almost certainly cannot name is the SHARE, and share is the entire information content of the list: whether your largest supplier is a comfortable third of your buying or a frightening half, whether your second could absorb the first, and how much of your office week is being spent administering accounts that carry almost nothing. A ranking withholds all of that behind a sort order, and it does it so smoothly that a shop can read the same list every quarter for years and learn nothing from it.

What the list is made of

• Population: suppliers you sent orders to inside the window.
• Ranking measure: committed spend, meaning orders TRANSMITTED to the supplier rather than invoices paid. Untransmitted drafts are not committed to anybody and do not belong in the ranking, which is the same split the open-purchase-order card makes.
• Presentation: a short top list, typically five rows, with the rest not shown at all.

Two things follow from committed spend being the measure. A supplier who forced you to cancel and re-source shows a SMALLER share than the business you actually tried to give them, because the cancelled orders left the record entirely, which is a defect that has its own article. And a single large order placed on the last day of the window lands in full, though nothing has arrived and nothing has been paid.

The window matters more than it looks. Use trailing 12 months for the concentration read, because a quarterly cut in any seasonal trade simply crowns whoever serves that season. Quarterly is for direction, not for level.

And there is a category of spend that never reaches the list at all. In most field shops some parts are bought at a counter on a card by whoever is nearest the job, with no order raised against anybody. That spend is invisible to a ranking built from purchase orders, it is often several percent of parts cost, and it is the spend that most clearly marks where your stocking gaps are. Before reading concentration at all, establish whether card and counter purchases are inside your ranking or outside it. If they are outside, what you are holding is a ranking of your PLANNED buying, which is a different and considerably more flattering thing than your actual buying.

Turn the ranking into shares

Twenty-three suppliers in the trailing 12 months, and this shop's card and counter purchases sit outside the ranking, so every share below is a share of spend that went out on a purchase order rather than a share of parts cost. The five that get shown, as shares of total committed spend, with each one's size against the next:

The top five hold 83 percent between them. The remaining 18 suppliers hold 17 percent, and that 17 splits again: 7 of them carry 13 percent between them, and 11 carry 4 percent.

That is the same list the ranking showed, and it now answers three separate questions with three separate denominators. Keep them separate. Mixing them is the specific way this analysis goes wrong.

Read one: dependence, which is an absorption question

The top supplier holds 41 percent of committed spend. The plain version of that is that in the week they cannot supply, 41 percent of the shop's parts flow has nowhere to go.

The useful version is an absorption test rather than a percentage. Rank 2 currently handles 19 percent. Covering rank 1 would put them at 19 plus 41, which is 60 percent of the shop's spend, or 3.2 times the volume they handle today. That is not a phone call. It is a different account, with a different credit limit, different terms, and above all a different stocking profile, because rank 2 stocks to your current pattern of buying and not to rank 1's.

So the question is never "do I have a second supplier". It is "can my second supplier take roughly three times their current volume from me inside my promise window", and for most shops the honest answer for the first two or three weeks is no.

The gate, and it runs against this example. If your top supplier holds more than about a third of committed spend, the switch plan gets written down: who takes each part family, what their lead time is, what your credit limit there is today, and who makes the call. At 41 percent this shop is over that line and does not have one. Treat the one-third figure as a starting point to tune rather than a standard, because there is no defensible published concentration benchmark for a shop this size, and the real gate is the absorption test above, which is operational rather than statistical.

None of this argues against concentrating your spend on purpose. Deliberate concentration is how a small shop buys leverage, and a sibling article covers how to do it. The two sit together perfectly well: concentrate, and hold a written switch plan as the price of concentrating.

Two suppliers can turn out to be one supplier. The absorption test assumes rank 2 is independent of rank 1, and frequently they are not. Two distributors carrying the same brand out of the same regional warehouse fail in the same week, so a switch plan naming one as the backup for the other resolves to nothing on the day it is needed. Finding out costs one question at each counter about where a given line comes from, and the historical version is cheaper still: think back to the last allocation or shortage that hurt you, and check whether both of them were short of the same thing at the same time. Where they were, count them as one row when you read your concentration, which for this shop puts the combined top position at 41 plus 19, or 60 percent, rather than 41.

Read two: leverage, where the denominator changes under you

This is the read most likely to go wrong, and it goes wrong in one specific way. 41 percent of YOUR spend is not 41 percent of THEIR business. Those are two ratios over two completely different denominators, and a shop that walks into a terms conversation holding the first while believing it has the second is negotiating with a number that is not a card.

Your leverage is your spend as a share of that supplier's revenue, and you cannot see it. What you can see are the revealed answers, which are better evidence than a percentage anyway:

• Do they hold stock for you by name, or do you take what is on the shelf like everyone else?
• Do you have a named inside-sales or counter contact who knows your work, or a queue?
• Is your discount standing and off list, or negotiated case by case each time?
• Will they open for you outside hours, and has that actually happened?

Each of those is a real answer to how much you matter there, and none of them costs anything to check.

The other half of this read is the one shops miss: leverage is won at the branch, not at the chain. A shop that is a rounding error to a national distributor can easily be a meaningful share of one branch's counter business, and the branch manager is who sets what gets stocked locally and who takes the after-hours call. Read your concentration at the branch level where the same chain serves you from two locations, because as a single chain-level row it will look like one large relationship when it is really two medium ones.

Read three: the tail, counted two ways

Eleven of the 23 suppliers, which is 48 percent of the account list by count, carry 4 percent of committed spend between them. Naming both bases matters here, because the two shares point opposite ways and the cost follows the count, not the money.

The administrative cost of a supplier is per-supplier: an account to open, terms to agree, a login, an invoice to match against a receipt, a statement to reconcile, a tax form at year end, and somebody to chase when any of those goes wrong. At even 15 minutes a month each in matching and reconciliation, 11 accounts is 2.75 hours a month, which is about 33 hours a year, close to a full working week of office time spent administering 4 percent of the spend.

That looks like an obvious cut, and cutting it blindly is the single worst thing this analysis can cause. Some tail suppliers are there because they hold the one part nobody else carries, and their share is low precisely BECAUSE that part is rare, not because the relationship is unimportant. A spend-ranked cull deletes exactly those, and the shop finds out on a Saturday.

So the tail gets sorted by something other than spend before any account is closed, and the test in the next section is what does the sorting.

The single-source test, run once a year

This is the exposure the ranked list cannot show at any level of detail, because it is not about suppliers at all. It is about parts.

1. Pull 12 months of parts ISSUED to jobs, not parts purchased. Issued is what the work actually consumed. Purchased includes things that are still sitting on a shelf and tells you what somebody bought, not what you depend on.
2. For each part, count the DISTINCT suppliers it came from in that period. Exactly one is your exposure list. Two or more, from genuinely independent sources, is fine.
3. Sort the exposure list by the number of JOBS the part appeared on, not by its spend. A common fitting used on 40 jobs is a far larger exposure than an assembly used twice, and sorting by spend puts them in the wrong order. This is the step that saves the tail supplier the previous section was about to cut.
4. For each of the top rows, write down two facts: the named substitute, and where it comes from. Not "we would find something", which is the answer that has never once been true at eight in the morning with a van already loaded. A part number and a supplier.
5. Where nobody can write the substitute, that IS the finding. Order a small quantity of the alternative and put it through one job before you need it, because a substitute that has never been fitted is a hypothesis.
6. Re-run annually, and out of cycle whenever a supplier is acquired, changes ownership, or drops a line. Those three events are what turn a two-source part into a single-source part overnight, and nothing in your own records will announce it.

Run that test against the shop above and the exposures almost never sit with rank 1. They sit in the tail, at a supplier holding 1 percent of spend and 100 percent of one part that goes on forty jobs a year. That is the whole argument for converting a ranking into shares and then looking past both: the list orders your suppliers by how much you spend, and your actual fragility is ordered by something else entirely.

References

• See related: Building Supplier Leverage as a Small Shop, for how deliberate concentration is used
• See related: Choosing a Primary Supplier, for the yearly primary-supplier review this sits alongside
• See related: Purchase Order Fulfilment Rate and the Orders It Never Counts, for why cancellations shrink a supplier's apparent share
• See related: Open Purchase Order Value Is a Commitment, Not a Cost, for what committed spend includes