Setting Up a Retail POS in Tanzania — Complete Guide

Choosing POS software in Tanzania: the questions that actually decide it

Most POS comparisons rank features that every product has. For a Tanzanian retailer, the decision usually comes down to four things that global feature lists barely mention.

1. Is it a TRA-certified fiscal solution, or does it bolt onto one? A POS that prints a sales slip and then requires a cashier to re-key the sale into a separate fiscal device has doubled your work and created a permanent source of mismatch between your sales ledger and your fiscal record. Certified end-to-end fiscalisation is not a premium feature here; it is the baseline.
2. Does it genuinely work offline? Not "cloud with a cache" — actually complete a sale, print a receipt and queue the fiscal submission with no connection at all, then reconcile cleanly on reconnection.
3. Does it handle mobile money as a tender type, per network, so the end-of-day reconciliation separates M-Pesa from Tigo Pesa from Airtel Money from cash rather than lumping them as "other"?
4. Can it price and sell what you actually sell? Loose rice by weight, sugar by scoop, unbarcoded local goods, a crate that breaks into bottles, a carton that breaks into pieces. A supermarket POS designed for pre-barcoded packaged goods falls over in a Tanzanian duka within a day.

Before you shortlist anything, establish whether the law already requires you to fiscalise, because that single answer removes half the products from your list. Our EFD/VFD eligibility checker asks three questions: are you VAT-registered, has annual taxable turnover passed the VAT threshold of TZS 200,000,000, and does your line of business appear on TRA's mandatory list? General retail is not automatically on that list, but hardware and wholesale/distribution are, as are bars, hotels, fuel stations, pharmacies and private schools. Answer yes to any one and the clock starts, with a monthly penalty thereafter that scales with your turnover.

EFD or VFD — pick the right device for your counter

An EFD is a physical fiscal printer sitting at the till, which suits a retail shop with a fixed counter and a cash drawer. A VFD is a software fiscal device, which suits invoicing, online sales, service businesses and mobile vendors. Many Tanzanian businesses run both: an EFD at the front counter for walk-in retail and a VFD for wholesale invoices and deliveries. Decide this before you buy hardware, because it changes what you buy.

A 30-day go-live plan

Days 1–5 — get your registration documents in order. TIN, VRN if you are VAT-registered, business licence, and the trading name exactly as registered. Fiscal receipts carry these details, and a mismatch between what TRA holds and what your POS prints will surface at the worst moment. Our VFD compliance checklist lists what you need in hand; our business licence fee tool covers the council side, including the 31 March renewal date and the 25% late penalty.

Days 6–15 — build the product file. This is the phase that decides whether the project succeeds, and it is the phase every retailer under-resources. Work item by item:

• One SKU per sellable unit. If you sell a carton of soap and also single bars, that is two sellable units with a defined conversion between them. Getting this wrong means your stock figure is wrong from day one and never recovers.
• Decide the barcode strategy per category. Manufactured goods carry a barcode — scan it and use it. Unbarcoded local goods, loose grains and produce need either your own printed labels or a well-designed quick-key layout on the till. Do not plan to "add barcodes later"; the cashier will invent workarounds and you will inherit them.
• Set the correct tax code per item, not a shop-wide default. Standard-rated lines carry VAT at 18% — use the VAT calculator to move between the VAT-inclusive shelf price customers expect and the net figure your margins are built on. Items with different treatment must be flagged individually, or every receipt you issue is wrong.
• Enter cost as landed cost, not invoice price, for anything imported. Duty, import VAT, the statutory levies charged on CIF, the clearing agent and the transport inland all belong in the item's cost — our customs duty calculator itemises the tax side of it. Costing at invoice price is the most common way a Tanzanian importer discovers, a year later, that a whole category was being sold below cost.
• Set the selling price deliberately. Check each line with the markup vs margin calculator — a 50% markup is a 33.3% margin, and the two get confused constantly.

Days 16–20 — hardware and layout. Till or tablet, receipt printer, scanner, cash drawer, and — critically for Tanzania — a UPS on the till and the router. A power cut mid-transaction on an unprotected till is how you corrupt a day's data. If you sell by weight, confirm the scale actually talks to your POS before you commit; a scale that requires the cashier to read a display and key a number in is a permanent source of shrinkage.

Days 21–25 — count the opening stock. Close, or count after hours. Count in teams of two, one counting and one recording, and count by physical location rather than by product list so nothing is missed. Enter the count as an opening stock document with a date and a signature, not as a series of ad-hoc adjustments. Everything you will ever say about shrinkage depends on this number being real.

Days 26–28 — users, permissions and training. Every cashier gets their own login. Shared logins destroy accountability: with one "cashier" account, a till shortage belongs to nobody. Set who can void a line, apply a discount, change a price, or process a return — and set it before go-live, because permissions granted in a rush during week one never get tightened afterwards.

Days 29–30 — parallel run, then switch. Run the new POS alongside your existing method for two full trading days including a Saturday. Compare totals daily. Go live only when they agree.

Week one: shift discipline and the numbers that prove it works

The operating routine matters more than the software. Issue a counted float at the start of each shift, signed for. At handover, run an X report — an interim, non-fiscal summary that does not close the day — and reconcile the drawer against it per cashier. At close, run the Z report, which is the fiscal end-of-day record TRA recognises, and reconcile in three separate streams: cash, mobile money by network, and card. Never net them together; a cash shortage hiding under a mobile-money surplus is invisible in a single combined figure.

Expect variance in week one and treat it as information rather than failure. Persistent shortage on one cashier is a training or integrity issue. Variance spread evenly across everyone is usually a product-file problem — a wrong price, a wrong pack size, a missing conversion. Fix the file, not the person.

The mistakes that sink retail POS projects here

• Going live during peak trade. Never go live in the week before Eid, Christmas or a school reopening. Choose your quietest fortnight.
• Skipping the opening count "because we'll catch up". You will not, and every stock report for the next year is fiction.
• Not testing offline behaviour. Pull the router during the parallel run, keep trading for twenty minutes, then reconnect and check that each of those sales lands in the fiscal totals once and only once. Duplicated or dropped offline sales corrupt your stock ledger and your tax position together.
• Leaving mobile money on a personal number. Collections into a personal wallet cannot be reconciled to a till and mix business and personal funds.
• Buying per-user pricing without counting the users. Cheap per seat becomes expensive once the storekeeper, the supervisor and the second branch are added.

Once the till is stable, the next wins are stock discipline and reordering. Our reorder point calculator sets the level at which each SKU should trigger a purchase order, and our guide to inventory management best practices covers counting cadence and shrinkage control. For the product itself, see the Tawala POS module and Tawala for retail.