South Africa's on-demand grocery market has passed the point where the app is the hard part. Checkers Sixty60 generated R11.9 billion in sales in the six months to December 2025, up 34.6%, and now accounts for more than 10% of Shoprite's South African supermarket turnover. It runs from roughly 875 stores with close to 10,000 Pingo drivers. Pick n Pay's asap! grew about 40% year on year in the first half of its 2026 financial year, operating from more than 620 stores with over 2,500 drivers.
The May 2026 South Africa Social Commerce Market Report put the same trend across the field, with Sixty60, asap!, Woolies Dash, SPAR2U and Takealot all expanding delivery and pickup. The demand is proven. The question is whether the delivery operation can hold its cost and service as volume climbs.
The cost sits in the last mile
Store-based picking and one-hour promises put strain in three places. Drivers have to be rostered against demand that swings by hour and by suburb. Routes have to stay dense enough that cost per drop does not creep up as the network widens. And every late or failed delivery carries a refund, a redelivery and a customer who tells others.
At 10,000 drivers, a small change in drops per driver per shift moves millions of rand a year. This is where the economics of on-demand grocery are won or lost, and it is a coordination problem before it is a technology problem.
What delivery orchestration does
Delivery orchestration software has grown up to handle exactly this layer. Platforms such as FarEye combine route planning, control-tower visibility across drivers and carriers, driver and roster management, digital proof of delivery, and returns into one system. FarEye reports more than 150 customers across 30 countries and positions itself for retail, e-commerce and 3PL teams running complex last mile at scale.
The value is in the coordination. Intelligent routing raises drops per route. Real-time visibility lets a control room re-assign a delivery before it fails rather than refund it after. Roster tools match driver supply to the demand curve instead of a flat shift pattern. For an operator adding stores and drivers every quarter, that coordination is what keeps cost per drop flat while volume rises.
The operating model comes first
Software alone does not fix a last-mile network. The operating model underneath it decides whether the software pays back. Which stores serve which zones. Where a dark store beats in-store picking. How many drivers a suburb needs at 17:00 on a Friday against 10:00 on a Tuesday. What service promise the density in a given area can actually support.
Those are design decisions, and they set the ceiling on what any orchestration platform can deliver. Get the network design wrong and the best routing software still runs half-empty vehicles.
Where OptiChain sits
The grocery delivery race in South Africa is now a last-mile operations race. The retailers growing fastest are the ones whose delivery model scales without cost per drop climbing with it.
OptiChain Solutions works with South African retailers and 3PLs on the last-mile operating model: zone and store-to-customer design, driver capacity planning against real demand curves, and the business case for orchestration technology before it is bought. The market has proven people will order. The margin question is whether the delivery network holds its cost as they do.