Every number we use, and where it came from
We are building for a market most technology commentary gets wrong by a factor of ten. So this page publishes the figures, the sources, and — just as important — which of our own numbers are still assumptions.
The market is large, formal and measured
South Africa's independent trade is not a rounding error. It is a quarter-trillion-rand channel with formal data — which is why we use published sources instead of anecdotes.
The wholesaler is the load-bearing node
95% of 118 surveyed independent traders (Ipsos, June 2024) buy their stock from wholesalers.
This single figure decides our architecture. If the wholesaler is the node, then the wholesaler must be observable — because a retailer can only be served what the wholesaler can see.
It also settles a common design argument: we onboard retailers through wholesalers, not around them. Bypassing the node would mean bypassing the only party with real supply information.
What this means in the product
- Wholesaler catalogue is the single source of product truth
- Retailers see stock their wholesaler has actually confirmed
- Supplier price-list import is a first-class wholesaler workflow
- Retailer onboarding runs through wholesaler relationships
The Spaza Shop Support Fund, and where it leaks
The DSBD Spaza Shop Support Fund is the clearest measurable demand signal in this market — and the clearest evidence that compliance is an access problem, not an administrative one.
| Stage | Businesses | What it tells us |
|---|---|---|
| Registered | ~82,000 | Demand to be formal exists at scale |
| Verified as real businesses | 44,696 | Registration alone proves very little |
| Reached a trading permit or licence | ~15,000 | The funnel breaks here — municipal licensing is the bottleneck |
| Approved for funding | 2,369 | ~R179.6m released (SEDFA R79.6m + NEF R99.9m) |
Terms
- R500 million fund size
- R100,000 per applicant, split 50% grant / 50% loan
- Stock capped at R40,000; assets at R50,000; non-financial at R10,000
- Requires bulk buying, a POS system, digital literacy and credit health
Why those criteria are our roadmap
All four criteria are software-addressable: bulk buying is a catalogue function, POS is the core transaction, digital literacy is adoption support, and credit health is a track record we create by having real transaction history. This is why the compliance tracker is a product module and not an afterthought.
The costs we are actually solving
Connectivity is affordable, not absent
GSMA data: coverage 92.2%, mobile internet penetration 62.15%, data affordability 54.2 per 100, handset affordability 43 per 100.
The barrier is cost and handsets — not towers. That is a design constraint, not an excuse.
Load shedding is a revenue event
Nedbank/TEA research: 64% of informal traders have stopped trading at some point, 66% shed jobs, and average monthly revenue loss exceeds R11,000.
Stock-outs compound
Roughly 8% out-of-stock, equating to about 4% of annual turnover. Around 27% of shoppers defect after a single stock-out.
Transport dominates logistics cost
Transport is roughly 59% of South Africa's logistics cost, at about R300–R350 per township collection trip.
Funding needs records, not estimates
Applicants are assessed on stock and sales records. A notebook is the single most common failure point — and the easiest to fix with a POS that works offline.
WhatsApp is the incumbent system
It is also unsearchable, unauditable, lost with a lost phone, and impossible to run analytics on. Replacing it means meeting it on its own terms — which is why messaging is built into the app.
Which of our numbers are assumptions
A market this size attracts a lot of invented precision. Here is the line we draw.
- Published — figures with a named external source, shown on this page
- Live — measured from our own running platform
- Derived — arithmetic on published figures, with the formula shown
- Assumption — our judgement, labelled as such, never dressed as data
Revenue trajectory
Our five-year revenue projection runs from roughly R22 million in year one to about R965 million in year five, on a blended take rate rising from approximately 3.7% to 4.0%.
These are management projections derived from an operating model. They are not audited results, not a forecast of return, and not a representation of achieved revenue.
Break-even in the model is roughly 821 active retailers, or around 631 at a 30% higher average basket. Both are model outputs, not operating targets.
Want the underlying pack?
We share the full market register, operating model and pilot arithmetic with serious counterparties.