Practical supply-chain digitisation for East and Southern African SMEs

Most advice about supply-chain digitisation for smaller African firms assumes the pressure comes from large buyers. Get your systems in order, the argument runs, or the supermarket chain will drop you. That is not what is actually happening in East and Southern Africa.

The deadline is being set by revenue authorities. Kenya, Zambia, Mauritius and Uganda have all made structured electronic invoicing compulsory for progressively smaller firms, and buyers enforce it second-hand because a non-compliant supplier costs them money at their own tax filing. That changes both the timetable and the order of work.

Start with the record, not the software

The first artefact is not a platform. It is a current, shared record of what you hold, what you owe and what is owed to you.

This is more than intuition. World Bank enterprise research across African microenterprises finds that point-of-sale and inventory-control software is one of a short list of technology uses significantly associated with higher employment, alongside accounting software and online banking. It is one of the few places where the link between a tool and a business outcome shows up in the data. The qualifier matters: that sample runs across seven sub-Saharan countries and is weighted heavily towards informal micro-enterprises, and the finding is a conditional correlation rather than proof that the software caused the hiring.

The honest counterweight belongs in the same paragraph. Among African microenterprises that do not use internet-supported technologies, more than seven in ten give no perceived need as the reason, rather than cost or availability. At the smallest end of the market the binding constraint is perceived usefulness, not availability. Anyone selling digitisation into that segment is arguing against a rational assessment, not against ignorance.

The payment rail is already digital, the paperwork is not

The acceptance layer is solved and operating at real scale. Safaricom’s Kenyan merchant base reached 3.13 million at 31 March 2026, up 71 per cent year on year, driven mainly by Pochi la Biashara reaching 2.09 million informal traders. In South Africa, Yoco reports more than 200,000 independent businesses processing 174 million transactions in a year.

Regionally the numbers are larger still. More than 2 trillion dollars flowed through mobile money globally in 2025, double the first trillion in only four years, and the GSMA reports that most new registered and active accounts came from sub-Saharan Africa. The same GSMA report puts the sub-Saharan share at roughly 1.4 trillion dollars, cited here via accessible coverage of it because the report page itself blocks automated retrieval. We covered the wider shift in fintech disruption in 2026.

Here is the gap that matters. A trader can accept digital payment from several thousand customers and still have no structured record of what was bought, from whom, at what price, or what remains on the shelf. Payment acceptance is not supply-chain visibility. The rail is digital and the paperwork is not, and it is the paperwork that revenue authorities have now started to require.

The deadline is being set by revenue authorities, not customers

Four countries in the region have already made structured electronic invoicing compulsory, and the thresholds keep descending toward smaller firms.

The enforcement mechanism is the part worth understanding, because it explains why buyers appear to be driving this when they are not. In Uganda, a buyer who accepts a non-compliant invoice loses both the input VAT credit and the expense deduction. In Kenya, a buyer purchasing from a micro-supplier has to raise the invoice themselves. In both cases the cost of your non-compliance lands on your customer, which is why your customer starts asking.

Mauritius is the clearest illustration of direction. The threshold descends from MUR 100m to MUR 80m in June 2026 and MUR 40m in September 2026. A firm comfortably outside the scope in 2024 can be inside it within two years without changing anything about its own business.

What large buyers actually ask for

Genuine private-buyer entry conditions exist, but they are narrower and more specific than the general warnings suggest.

Barcodes that validate. GS1 South Africa states that barcode numbers issued by anyone other than an official GS1 member are often rejected by Checkers, Takealot, Food Lovers Market and Google, and that retailers run supplier submissions through its validation tool. Many require a barcode test report as a listing step. This is a real gate, and it is cheap to clear if you know about it before you print packaging.

Plot-level geolocation for certain exports. The EU Deforestation Regulation requires a due diligence statement with plot-level geolocation from 30 December 2026 for large and medium operators, and 30 June 2027 for micro and small ones, covering coffee, cocoa, rubber, palm oil, soya, cattle and wood. For a Kenyan, Ugandan or Tanzanian coffee exporter this is a data requirement that cannot be retrofitted from order history. The plots have to have been recorded.

One correction worth making plainly, because it appears in a great deal of advice on this subject. We found no regional evidence that supermarket or manufacturing buyers in East or Southern Africa systematically require electronic advance shipping notices or automated purchase-order responses from SME suppliers. If someone tells you that is the entry condition, ask them for the buyer and the document.

The constraints are cost and perceived need, in that order

An entry-level 2 GB mobile data basket cost a median 4.2 per cent of monthly income per capita across Africa in 2024, against a 2 per cent affordability target and 1.1 per cent globally. The average conceals the real problem. Even in Kenya, where the basket sits below 2 per cent of average income, only around 40 per cent of the population can buy it for under 2 per cent of their own income, a figure derived from the income distribution rather than the national average.

Coverage is not the binding constraint either. At the end of 2021, roughly 84 per cent of the sub-Saharan population was within 3G range while about 22 per cent actually used mobile internet. More recent GSMA figures put coverage a little higher and usage in the mid-twenties, so the gap has narrowed without closing. The gap between coverage and use is where digitisation projects fail, and it is made of cost, device access and perceived usefulness rather than infrastructure.

This is why sequencing matters more than ambition. A system that requires a supervisor to be online continuously will not survive contact with a business whose data spend is a considered monthly decision.

A sequence that survives the constraints

Work in this order. Establish the stock and purchase record first, on whatever tool the team will actually use daily. Get compliant with your revenue authority’s invoicing system second, because that deadline is fixed and the penalty is arithmetic. Add buyer-specific requirements third, and only the ones your named buyers actually ask for. Automate last, once the record is trustworthy enough to be worth automating.

That order is deliberately unexciting, and it is the same discipline that separates firms which get value from technology pilots from those which do not, a pattern we set out in moving agentic AI from pilot to production. The failure mode is identical in both cases, which is buying the sophisticated layer before the boring one underneath it works.

Key takeaways

• The digitisation deadline in East and Southern Africa is being set by revenue authorities, not by large buyers. Kenya, Zambia, Mauritius and Uganda already mandate structured electronic invoicing.
• Buyers appear to drive it because a non-compliant supplier costs them the input VAT credit and the expense deduction.
• Mauritius thresholds descend to MUR 80m in June 2026 and MUR 40m in September 2026, so firms outside scope today can be inside it shortly.
• Payment acceptance is not supply-chain visibility. A trader can take digital payment from thousands of customers and still hold no structured record of stock.
• Verified buyer requirements are narrower than commonly claimed, principally GS1 barcode validation and EUDR plot-level geolocation for certain exports.
• The binding constraints are data cost and perceived usefulness, not network coverage.

Frequently asked questions

Which comes first, the accounting system or the inventory record?

The record. An accounting system fed by an unreliable stock count produces confident numbers that are wrong. Establish what you hold and what you owe on a shared record the team updates daily, then connect it to accounting. The evidence associating software with employment growth in African microenterprises covers point-of-sale and inventory control alongside accounting, not instead of it.

We are below the VAT threshold. Does electronic invoicing still apply to us?

In Kenya, yes. The eTIMS obligation covers all persons in business regardless of VAT registration, and where your turnover is KES 5 million or less the buyer raises the invoice on your behalf. In Uganda, EFRIS now applies across twelve named sectors regardless of VAT registration. In Zambia and Mauritius the obligation is currently tied to VAT registration or turnover thresholds, but the Mauritian thresholds are descending on a published schedule.

How much should a small firm expect to spend on this?

Less than most vendors imply, if the sequence is right. Revenue authorities provide the invoicing tools, and Zambia supplies a free Android application for Turnover Tax taxpayers. The recurring cost that actually binds is connectivity, where the regional data basket sits above the affordability target for most of the population.

Do we need to be ready for the EU Deforestation Regulation now?

If you export coffee, cocoa, rubber, palm oil, soya, cattle or wood into the EU, the timing matters more than the paperwork. Obligations apply from 30 December 2026 for large and medium operators and 30 June 2027 for micro and small ones. The geolocation data has to have been recorded at the plot, so this is one requirement that genuinely cannot be assembled retrospectively.

Our buyer has not asked for any of this. Should we still act?

Treat the tax mandate as fixed and the buyer requirement as conditional. The invoicing deadline applies whether or not anyone asks. Buyer-specific requirements like GS1 barcode validation are worth clearing before you commit to packaging or a listing application, but there is no case for building capability against requirements no named customer has stated.

Sources

• Kenya Revenue Authority, eTIMS, what is eTIMS. kra.go.ke
• Zambia Revenue Authority, Smart Invoice. zra.org.zm
• Mauritius Revenue Authority, e-invoicing. mra.mu
• PwC Uganda, EFRIS compliance. pwc.com/ug
• KPMG, South Africa confirms multi-year e-invoicing reform, February 2026. kpmg.com
• Safaricom, FY26 results, 7 May 2026. safaricom.co.ke
• Yoco, uncovering the data. yoco.com
• GSMA, Mobile Money accounted for $2 trillion in transactions in 2025, press release, 24 March 2026. gsma.com
• Connecting Africa (Informa), reporting the GSMA State of the Industry Report on Mobile Money 2026, 26 March 2026, for the sub-Saharan regional figure. connectingafrica.com
• European Commission, EUDR implementation timeline. trade.ec.europa.eu
• GS1 South Africa, barcode verification. gs1za.org
• ITU, ICT price trends 2024. itu.int
• World Bank, digital technology adoption among African firms. worldbank.org