Haute Lumière

Commerce · II.08 · MMXXVI · daylight

La Bourse  /  Volume II  /  Nº II.08  /  Workbook — the executive

WORKBOOK — THE CORPORATE EXECUTIVE

Chapter II.08 · Networks, Trust, and Topology

For the person with a P&L, a supplier base and a signature limit. The argument here is not that trust is good. It is that your counterparty network is an unmanaged asset with a measurable cost of capital, and that nobody in the building currently owns it.

---

THE PREMISE, STATED COMMERCIALLY

Your firm buys assurance on every transaction. It buys it two ways — through relationship, which is a fixed cost paid in time, or through contracting, which is a variable cost paid per order. Almost no firm knows the split, and almost every firm has them the wrong way round: relational overhead lavished on low-volume suppliers, and high-volume strategic partners handled on templates by someone three levels down.

The prize is not soft. Dyer and Chu surveyed 344 supplier–automaker relationships across three countries and found procurement transaction costs roughly five times higher for the least trustworthy buyers than for the most trustworthy. Wallis and North measured the whole American transaction sector and found it rising from about 25 percent of GNP in 1870 to over 45 percent by 1970. Assurance is the largest line item nobody itemises.

This workbook takes ninety days and produces four artifacts: a two-curve table of your counterparty base, a named trust core chosen by transaction volume, a published entry ladder, and a facility memo with one inequality on the front page.

---

PART ONE — DISCOVERY

Days 1–30: measure what you already have

Exercise 1.1 — The two-curve table (the whole of week one)

One row per counterparty. Five columns:

The fourth column is the one nobody has. Build it once from a bottom-up estimate: template retrieval, credit check, purchase-order match, goods-inward inspection, invoice matching, collections provision. In a well-run firm it lands somewhere near 250 per transaction, which on a 10,000 order is 2.5 percent. Use your own figure; the method matters more than the constant.

The fifth column is the one people lie about. Get it from calendars, not from memory.

Exercise 1.2 — The mismatch scan (2 hours)

Sort the table by relationship hours and read the transaction counts beside them. Every firm that has done this finds the same two faults: substantial relational time spent on counterparties transacting four times a year, and high-volume partners with no named relationship owner at all.

Correcting the mismatch is free and it usually funds the rest of the programme. You are not spending more; you are spending the same hours on the other rows.

Exercise 1.3 — Find your positive core (half a day, with procurement)

Ask directly, in a room: which counterparty relationship saves us real money every month, and who built it? You are looking for the relationships that are already outperforming — the supplier who flags a design fault before it becomes a recall, the one who holds stock for you without being asked, the one who has never once invoiced in error.

Cost one of them properly, including the counterfactual. What would a transactional relationship with that supplier have cost this year in expedite fees, quality escapes and rework?

Exercise 1.4 — Count the door (1 hour)

How many counterparties has the firm admitted to its top tier in the last three years? If the answer is zero, you are running a closed network, and closure has a price you are already paying.

---

PART TWO — THE ARITHMETIC

Days 31–45: price both topologies

Exercise 2.1 — The crossover, on your own numbers

  trust, per member per year      C_t(n) = 150 × (n − 1)
  contract, per member per year   C_c(T) = 250 × T
  crossover                       n = 1 + 1.667 T

Substitute your own cost of an hour and your own contracting cost. The ratio is what matters, not the constants.

What the table is actually telling you: the trust curve depends on membership and the contract curve on transactions. A wide, warm, low-volume network is the most expensive structure you can operate. A 300-member association doing eleven pieces of business a year with itself is carrying 44,850 of maintenance per member against 2,750 of contracting.

Exercise 2.2 — The price of your own closure

  E[best of k] = 0.9 + 0.2/(k + 1)
  open to 40 qualified suppliers   0.904878
  closed to  8 network insiders    0.922222
  gap                              1.7344 % of purchase spend
  on 20,000,000 of spend           346,883 per year

Estimate k — genuinely qualified counterparties in the category — and k' — how many you actually invite. Compute the gap. This figure belongs in the same paper as the trust dividend, because it is the cost of the same decision seen from the other side.

Then compute what closure costs the people outside. On the eBay reputation experiment, an established identity earned 8.1 percent more than the same goods sold under a new one: a competent newcomer with 500,000 of revenue pays 40,500 a year for being new. The ratio of outsider cost to insider cost is about 4.67 to one.

Exercise 2.3 — The honest negative, in your own board paper

Write two sentences you will not be tempted to cut. The trust-and-growth literature gives 0.8 percentage points of growth per ten points of trust in Knack and Keefer's 29-country sample and 0.67 in Zak and Knack's; the effect did not survive extreme bounds analysis in the extended sample, and the strongest causal evidence is Algan and Cahuc's inherited-trust instrument.

Put that in the paper. A board that finds the weakness itself will discount the whole case; a board that is handed the weakness will discuss the proposal.

Exercise 2.4 — The concentration check

Uzzi found firm failure risk to be U-shaped in the share of embedded ties. Run the test: for every critical input, how many qualified alternatives have been kept live in the last eighteen months? The answer is very often zero, and it was produced by the same relationships you are about to invest in further. Price the re-qualification lead time. That is your concentration exposure, and it belongs in the risk register before the facility is signed.

---

PART THREE — DESIGN

Days 46–60: the instrument

The structure: an open trusted-counterparty facility.

Three tiers.

• Standard. Template terms, standard payment, full goods-inward inspection.
• Preferred. Fifteen additional days of payment terms, reduced inspection, earned on twelve consecutive conforming deliveries.
• Core. Shared forecasting, open-book costing, joint improvement targets. Entered from Preferred on measured performance.

Two Core places are reserved each year for counterparties with no prior relationship to the firm. Publish that. An unpublished ramp is not a ramp.

The worked book.

  annual purchase spend                        20,000,000
  suppliers                                            40
  procurement transaction cost at 2.5%            500,000

  top 12 suppliers, 60% of spend               12,000,000
  rate falls 2.5% -> 0.8%, a fall of 1.7 points
    saving 12,000,000 × 1.7%                      204,000

  maintenance 12 × 40 h × 75                       36,000
  15 days terms on 12,000,000 = 493,151 at 9%      44,384
  expected credit loss 0.5% × 493,151               2,466
  ----------------------------------------------------------
  total cost                                       82,849
  NET ANNUAL BENEFIT                              121,151
  return on incremental working capital              24.6 %
  gross of the credit provision                      34.1 %

The balance-sheet treatment. The working-capital leg is an extension of trade payables' duration financed at the firm's cost of capital — treasury's language, in treasury's part of the pack. The transaction-cost saving lands in operating expense, not in a sustainability appendix, because it is not one. Where the facility funds joint improvement work that creates a tooling or process asset, capitalise on the usual test and disclose counterparty concentration, which the auditors will ask about and should.

The counterparty. Internal first: procurement and treasury, documented in a fortnight. The external version of this structure is supply-chain finance, and you will price it materially better after two years of ledgered supplier performance, because that ledger is the credit file.

The number on the front page.

     transaction-cost saving − relationship maintenance
  --------------------------------------------------------  >  WACC
      incremental working capital + verification cost

On the book above, (204,000 − 36,000) / 493,151 = 34.1 percent before the credit provision and 24.6 percent after it, against a 9 percent WACC. And the reserved places are worth up to 346,883 a year in supplier-selection option value, which means the open door is the profitable half of the design.

---

PART FOUR — DESTINY AND DELIGHT

Days 61–90: make it hold

Exercise 4.1 — Get it into the standing pack

One page, monthly: counterparty tier movements, transaction cost per order by tier, days of terms extended, conforming delivery rate, and the count of new entrants admitted. Anything reviewed monthly persists; anything reviewed by exception does not.

Exercise 4.2 — Attach compensation to supplier quality, not supplier comfort

Not necessarily much. An unpaid metric is a hobby. Make sure the metric is performance of the base, not satisfaction of the incumbents, or you have paid someone to keep the network closed.

Exercise 4.3 — Publish the ladder outside the building

On the supplier portal, in the tender pack, on the website. The commitment that survives your departure is the one an outsider can quote back at your successor.

Exercise 4.4 — Name the second owner

One person is a hobby; two is a practice. Recruit them by giving them the credit for the first result.

---

THE FAILURE MODES, NAMED

So you can see them coming

The core is chosen by seniority. The tell: the Core list has not changed in three years and matches the golf list. The fix is mechanical — rank by T and complexity, and let the arithmetic pick.

The ramp exists on paper. The tell: you cannot name the last counterparty who walked up it. The fix is the reserved number and a January date.

Maintenance hours rise and transaction cost does not fall. The tell is in the two columns side by side. Those hours are being spent on dispute management, not relationship building — the same hours doing the opposite job.

Admission moves from performance to introduction. This is the moment a trust network becomes a closed network and the arithmetic changes sign. The tell: the last three entrants were all introduced by the same two people.

Concentration is mistaken for partnership. The tell: no live alternatives in a critical category. Cap it by policy, not by intention.

---

THE NINETY DAYS ON ONE PAGE

---

BOARD PAPER TEMPLATE

One page. In this order.

1. The proposal in one sentence. Move the twelve highest-volume counterparties to relational terms, publish an entry ladder, and reserve two places a year for new entrants.
2. The number. Net annual benefit 121,151; return on incremental working capital 24.6 percent against a 9 percent WACC.
3. What it costs. 82,849 a year, of which 44,384 is the carry on fifteen days of extended terms.
4. What it risks. Counterparty concentration, capped by policy at a stated share of category spend, with two live alternatives per critical input.
5. The honest weakness of the evidence. The macro trust-and-growth literature is contested; this proposal does not rest on it. It rests on the firm's own transaction-cost measurement, which is in the appendix.
6. What we are asking for. One signature and a standing pack line.

---

THE BRIDGE BUDGET

The commercial half nobody puts in the paper

Everything above is about the counterparties you have. The other half of the topology is the ones you have not met, and it is worth its own line.

Granovetter's finding is a commercial instruction, not a sociological one. Of jobs found through a personal contact in his Newton study, 83.4 percent came from someone the finder saw less than twice a week. The causal replication — five years of randomised variation across 20,000,000 people producing about 2,000,000,000 new ties and 600,000 jobs, one job per 3,333 ties — found moderately weak ties doing the most work and the strongest ties the least.

Read that as three budget lines.

In hiring. Set a stated proportion of every intake that arrives from outside the referral graph. Not as a gesture — as a search-quality instrument, on the same expected-best-of-k arithmetic you used on suppliers. Referral hiring is cheap and fast and it narrows k every year it runs unopposed, and the audit literature is unambiguous about what a narrowed k does to who gets seen: Pager, Western and Bonikowski found white testers receiving positive responses on about 31 percent of applications against 15 percent for identically credentialled Black testers, a ratio of about 2.07 to one, in a labour market where most vacancies travel along ties.

In business development. Count what proportion of your pipeline originates inside the existing client cluster. A cluster of fifteen returns zero newly reachable people per additional tie inside it; one bridge to a different cluster of fifteen returns fourteen. If every origination is intra-cluster, you have a pipeline that cannot grow faster than your existing clients do.

In your own diary. Name the hours. Defend them when the quarter gets hard, because they are the first thing cut and the last thing that would have paid.

---

THE SUPPLIER CONVERSATION, SCRIPTED

For the first call with a counterparty you are moving to Preferred. It takes about six minutes and it is worth rehearsing, because the usual version of this conversation promises warmth and delivers nothing.

We have measured what it costs us to transact with each of our suppliers. You are in the top twelve by volume and your delivery record is clean, so we are moving you to Preferred terms from the first of next month. That is fifteen additional days of payment terms and reduced goods-inward inspection.

There are two things I want to be straight about. The first is that this is earned on measured performance and it can be lost the same way — here is the measure, and you will see it monthly. The second is that we are not asking for a price concession in return. The saving is on our side of the transaction and we would rather have the reliability than the discount.

In eighteen months there is a Core tier that opens up shared forecasting and open-book work. Here is what it requires. Nobody is nominated into it.

Three things make that script work: the criterion is stated, the reversal is stated, and nothing is asked for in exchange. A trust dividend offered with a price demand attached is a negotiation, and it will be treated as one.

---

APPRECIATIVE QUESTIONS FOR YOUR LEADERSHIP TEAM

1. Which counterparty relationship saves this business real money every month, and who built it? Has anyone ever said so in a document that counts?
2. When did we last admit a supplier nobody here already knew, and what did they go on to do for us?
3. Where are we already doing open-book work well, and what made that possible?
4. If every counterparty could see exactly what the ladder into our top tier requires, who would start climbing it this year?
5. What would our procurement function do differently on a Tuesday if it held both the maintenance cost and the transaction saving for every supplier?
6. What is the first sign we would see if our supplier base had quietly closed, and who in this room would notice it first?