The Customer Comes Before the Job Count

Part 4 of 10 in the 10,000 People Economy series.

A community employment plan allocates 760 people to food and agro-processing. The number fits the model, the work is useful and the community certainly needs food.

But who buys the output?

How much do they buy? At what price? How long do they take to pay? What does the producer spend on ingredients, packaging, power, transport, equipment and wages before the payment arrives?

Until those questions are answered, 760 is a labour allocation. It is not an employment result.

Need Starts the Design

South African communities contain urgent needs. Housing must be repaired. Children need early learning. Older people need care. Water systems need maintenance. Businesses need bookkeeping, logistics and technical support.

Need explains why the work matters.

It does not identify the funding route.

A household may need a service it cannot afford. A municipality may have an obligation without an approved contract for the proposed provider. A business may express interest without placing an order. A customer may order and pay sixty days after the producer has already paid two months of wages.

Usefulness starts the design. The operating question is who accepts the output and how the cash eventually reaches the worker.

Demand Is More Than a Market-Size Slide

A market estimate can show that thousands of people buy a product. It does not prove that they will buy from this enterprise at this price.

A survey records what people say. A letter of intent records interest. A purchase order records a transaction. A contract creates obligations. A paid invoice proves that an accepted product became cash.

These are different evidence levels.

An early pilot does not need five years of trading history. It does need to state how weak or strong its customer evidence is and limit its recruitment and capital accordingly.

The 10,000 People Economy uses an evidence control called E15. For every enterprise it requires:

• a defined product or service;
• a customer or authorised public purpose;
• volumes and prices;
• payment terms and collections;
• payroll, hours and loaded wages; and
• the working-capital cycle.

Until that evidence is independently accepted, the role remains planned rather than demand-backed.

The Current Number Is Large Because the Economy Is Large

The model allocates 6,460 productive placements inside a community of 10,000 residents.

Supporting those placements requires more than the wage bill. The current planning account requires approximately:

• R927.4 million in annual loaded payroll;
• R2.011 billion in annual sales;
• R1.461 billion in break-even sales; and
• R385.2 million in working capital.

These are not forecasts. They are the current answer to a harder question: what commercial burden sits behind the employment target?

Under a combined customer-volume and collection shock, the supported placement number falls from 6,460 to about 5,791.

The list of jobs has not changed. The customers bought less or paid later. Employment capacity changed.

The customer therefore has to come before the employment claim.

Working Capital Is Part of the Job

Annual revenue can hide a business that fails in month two.

Employees and suppliers must often be paid before the customer pays. Inventory sits in storage. Construction work reaches milestones before invoices are approved. Training providers deliver learner-months before a funder releases payment.

Working capital carries the enterprise through this gap.

If it is missing, workers become the lender through late wages, suppliers become the lender through unpaid accounts, or the business stops producing.

A credible employment model therefore asks not only whether annual sales exceed annual costs. It asks whether cash is available on the days payroll and inputs are due.

Public Value Still Needs an Account

The customer rule should not be used to exclude work that markets price badly.

Care, safety, public-space maintenance and basic services may have an authorised public purpose instead of a private customer. The evidence is different, but it is not weaker by definition.

The programme must show:

• the lawful service obligation;
• the population or infrastructure served;
• the delivery standard;
• the budget, contract, tariff or cross-subsidy;
• the full delivery cost; and
• the evidence that the service was provided.

Calling the work socially valuable is not a funding mechanism. Calling it non-commercial is not a reason to leave it outside the economy.

Demand Can Be Built

Putting the customer first does not mean waiting for a perfect market.

Public procurement can create an initial market. Development finance can fund productive capability. Shared quality systems can help small firms reach larger buyers. Export support can connect a community to new purchasing power.

The discipline is that these interventions must have tests.

A pilot procurement contract should have a volume, price, quality requirement, review date and decision rule. A supported enterprise should show whether it is learning, collecting cash and becoming less dependent on the intervention.

Demand formation is legitimate economic work. It is not permission to count prospective customers as current jobs.

The Employment Number Should Have States

One headline number cannot describe an employment programme honestly.

The programme should report:

• planned placements;
• customer-supported placements;
• revenue-supported placements;
• working-capital-supported placements;
• people actually placed; and
• independently accepted employment.

The final claim should not exceed the weakest binding constraint.

In the current 10,000 People Economy evidence state, the demand-backed employment number is zero. No complete enterprise-by-enterprise E15 schedule has yet been independently accepted.

This does not prove that there are no customers. It proves that the evidence required for the public claim is not complete.

Build the Transaction That Carries the Worker

Employment planning should begin with useful work. It should not end there.

For each role, the programme must build the chain from output to customer, from customer to collection, and from collection to wages, inputs and reinvestment.

The job count is the result of that chain.

When the customer is missing, the programme has not yet found an employment outcome. It has found the next institution it must build.

Sources

• 10,000 People Economy Project. (2026). Internal Flows Model Specification v1.
• Kim, K. (2011). Ex-ante evaluation of a targeted job program. Economic Modelling, 28(6), 2683-2690.
• Mazzucato, M., and Penna, C. C. R. (2016). Beyond market failures. Journal of Economic Policy Reform, 19(4), 305-326.
• Rodrik, D. (2004). Industrial Policy for the Twenty-First Century.

---

Previous: When Money Circulates but the Economy Remains Dependent
Next: Training Should Be Part of Production