Corporate & Business · NPO & Social Impact · Procurement & Sourcing
The cheapest way to cut a per-meal rate is to serve less food
Which is exactly what the mandate ruled out. Catering had been bought locally for years across local caterers, retail channels, venue-linked providers and ad hoc payment routes — at materially different prices, menus and service standards.
- ~127,000
- meals in the reconstructed baseline
- 14.6%
- reduction on addressed spend
- 6.6%
- equivalent across the national baseline
- 23.8–32.7%
- two migrated sites, specification improved
Proof context: A nationally dispersed operation catering across nine provinces
The situation
Twenty-six locations, nine provinces, sixty-six payment routes
Catering had been bought locally for years. Across the network, branches used local caterers, retail channels, venue-linked providers and ad hoc payment routes, at materially different per-meal prices with different menus, delivery arrangements and service standards.
The mandate was explicitly two-sided: reduce the cost, and preserve or improve meal quality, quantity and operational practicality. In catering those pull against each other more directly than in most categories, because the cheapest way to lower a per-meal rate is to serve less food.
What KuTh did
Reconstruct, verify against reality, then normalise
- Rebuilt the national baseline. Approximately 3,700 transaction records, 66 supplier and payment-channel entries and 26 operating locations across nine provinces, reconciled through 33 project and cost codes.
- Checked the data against the branches. Branch-level operational interviews, because the recorded supplier was not always the supplier actually being used. Accounting data describes who was paid, not always who cooked.
- Normalised against specification, not price. Meal specification, menu breadth, beverage size, portioning method, special dietary handling, delivery, equipment and service were all compared before any rate was called cheaper.
- Tested consolidation only where geography allowed. Regional routes were built where branch density and supplier reach supported them. Local suppliers were retained where a single-provider model would have been impractical — no blanket national model was imposed.
- Converted the high-volume sites to contract. Site visits, menu comparisons and supplier negotiation turned selected locations into contracted or contract-supported commercial outcomes rather than informal arrangements.
Results
Each workstream, with its status and its service effect
Scroll table sideways →
| Workstream | Reduction | What happened to the service |
|---|---|---|
| Gauteng anchor site | 8.5% | Existing provider retained, rate renegotiated against high recurring volume (approximately 28,900 meals a year). |
| Gauteng migrated site A | 23.8% | Moved to a regional route with a broader buffet specification and a larger standard beverage. |
| Gauteng migrated site B | 32.7% | Same regional route; menu breadth and service model improved. |
| KwaZulu-Natal, three sites | 13.3% | Incumbent retained. Menu rationalisation and waste removal, with quality and quantity requirements preserved in a formal agreement. |
Combined across the finalised and contract-supported workstreams: approximately 14.6% of the addressed spend base, equal to approximately 6.6% of the reconstructed national baseline. Roughly 58,000 meals a year fall under the completed changes.
The lower price came with more food, not less
On the two migrated Gauteng sites the reduction coincided with a broader meal specification, a larger standard beverage and a move to self-service. The saving came from the supply route and the service model, not from shrinking the offer.
This matters more than the percentage. A catering cost reduction that quietly reduces the meal transfers the cost to the people being fed rather than removing it.
Two denominators, both published
14.6% measures the reduction against the spend actually addressed by the finalised workstreams. 6.6% measures the same benefit against the full reconstructed national catering baseline.
The difference is the regions still open at the time of reporting. A further material opportunity was modelled in one province and negotiations continued in several others; none of that is included in the figures above.
Supporting documents