How to Choose a POS System in Kenya
A point of sale system decides three things about your business: whether you can trade when the network fails, whether your stock figures are true, and whether you know which products actually make money. Choose wrong and you pay for it twice — once for the software, then again in shrinkage, re-keyed invoices and month-ends that take days.
This guide walks through the questions that matter for a Kenyan shop, restaurant, bar or pharmacy, and what to test in a trial account before you commit to any vendor.
What “POS system” means for a Kenyan business
A POS (point of sale) system is the software at the till plus the back office behind it: product records, stock, customers, payments, staff permissions and reports. In Kenya the phrase covers three broad shapes, and the differences matter more than the feature lists:
- Desktop systems tied to one machine. Usually sold with a PC or an all-in-one terminal. They work offline by design, but adding a second branch means another licence, another machine and a way to merge two sets of books.
- Hosted (cloud) systems. The software and the database live with the vendor; you pay a subscription and open the till in a browser. Adding a branch is a setting rather than a purchase — provided the till keeps working when the connection drops.
- Point of sale apps bolted onto accounting software. Strong on ledgers, weak at the counter: slow to scan, awkward with split payments, and rarely built for a cashier with a queue.
For most traders the decision is really about the middle option, and the honest test for it is a single question: what happens to my till when Safaricom or KPLC takes the internet down?
The offline test, first
Any cloud POS will claim it works offline. The claim hides a range of behaviour, from “you can look at last night’s data” to “ring up sales, print receipts, hold payments and sync when the network returns”. Only one of those keeps a shop trading.
Ask the vendor to demonstrate, in front of you, a sale rung up with the network cable unplugged — including the receipt printing and the payment. BiznaPOS runs the register as an installable web app (a PWA), so the browser holds its own queue of sales and payments on the device and replays them to the cloud when the connection returns. Nothing is re-keyed and the reports match the drawer.
Stock is the part that decides whether you trust the numbers
Most owners do not buy a POS to ring up faster. They buy it because the shelf and the screen disagree. Look for these specific behaviours:
- One ledger per warehouse. Sales, purchases, returns, transfers and counts must all move the same number. If each lives in its own table, variances are unrecoverable by design.
- Reservations. An unpaid quote or a held order should reduce available stock without removing physical stock, otherwise two customers are promised the same item.
- Transfers with a receiving step. A branch-to-branch movement needs an outgoing and an incoming entry, or the sending branch absorbs the loss.
- Consolidated stock counts. Counting three locations should be one exercise with one reconciliation, not three spreadsheets compared afterwards.
- Batches and expiry. Non-negotiable for pharmacies and chemists, and useful for any perishables: alerts before stock becomes unsellable rather than after.
Payments: M-Pesa, card, cash and the split
In Kenya a till that cannot take M-Pesa properly is a till with a queue problem. “Properly” means three things, and you should check each one:
- The customer gets the STK-push prompt on their phone from the sale screen, rather than the cashier dialling a separate till or a phone.
- The confirmation is recorded against that sale, so the daily M-Pesa total reconciles to the sales report without a manual match.
- Split payments work — part cash, part M-Pesa, part card, part store credit on one bill, with each leg visible on the receipt.
Store credit, loyalty points, deposits and gift cards belong in the same conversation. If a vendor quotes them as “add-on modules”, ask what the add-on costs per month, for how long, and whether the customer record survives the integration.
Multi-branch: one set of books or three
A second branch is where most systems quietly fail. What you want from a point of sale for multiple branches:
- Per-branch stock, per-branch pricing and per-branch staff permissions, so a cashier in Thika cannot see or change Mombasa’s cost prices.
- A group view that rolls sales and margins into one report the owner can open on a phone.
- Cash-up per till and per shift, with supervisor approval at close and a variance report that names the shift and the person.
If the answer is “we would run a separate account for each shop and export at month-end”, that is not a multi-branch system — it is one branch repeated.
Cost: what a POS system in Kenya really totals
The subscription is the smallest part. A desktop system typically totals a terminal or all-in-one machine, a licence, an installer’s visit, support contracts per branch and a hardware repair you fund when the printer board dies. A hosted system replaces most of that with a monthly fee and the devices you already own.
Compare the full plan and price breakdown before deciding; BiznaPOS is free on a single counter (Starter, up to 500 products), 49 US dollars a month on Pro for up to five stores and 149 on Enterprise, with no setup fee and no per-till licence.
Eight questions to ask any vendor, in the room
- Unplug the network. Can I still sell, print and take a payment? What syncs afterwards?
- Show me a stock transfer between two branches and the report that proves it landed.
- What exactly is metered — products, stores, tills, users, or sales?
- Which features are in the base price and which are add-ons?
- Can I import my current catalogue from Excel with cost, price, barcode and opening stock?
- Can I export my sales, stock and customers, and does my data stay available if I leave?
- Who answers support, in which hours, and in which time zone?
- What breaks if I add a second till on the same counter?
The trial is the audit
Do not accept a demo on someone else’s data. Load your own top fifty products, price them, add your real staff users and run a busy hour. Then check three numbers: the cash-up against the drawer, the stock on the shelf against the screen, and the M-Pesa total against the statement. A system that agrees with all three on day two will agree with them in month twelve.
Every plan here starts with a 14-day trial and needs no card, so you can run that test on your own counter: create your shop.