Ask three vendors what a POS system costs in Kenya and you will get three monthly figures, all of them roughly similar, none of them the real number. The subscription is the one line everybody quotes because it is the one line everybody can win on. It is also, in most shops, somewhere between a quarter and a half of what the system actually costs you in year one.
This is the whole bill. Seven lines, in the order they arrive, with the Kenyan specifics that do not appear in a global pricing page — and a worked total at the end for a single shop and for a three-branch business.
Nothing here is an argument for buying one. It is the arithmetic you need in order to decide.
Line 1: The subscription — and the word "per"
The headline number is the easy part. What costs people money is the preposition next to it.
Read any quote for the word per. Per terminal, per till, per user, per cashier, per branch, per location. A system at KSh 3,000 "per terminal" in a supermarket running four lanes is KSh 12,000, and that is before the back office logs in. The same quote in a one-counter duka is KSh 3,000. Two shops, the same sticker, a four-fold difference in what you pay.
Then check whether the online store is inside the subscription or beside it. A POS vendor who charges separately for ecommerce is charging you twice for the same product catalogue, and you will maintain that catalogue twice as well — which is the expensive part, not the fee.
For reference, sell.ke's own plans run from KSh 1,499 a month for a single-counter shop to KSh 14,999 for a wholesaler or a chain, with no per-terminal and no per-cashier charge, and the online store included from the Basic tier up. The full breakdown is on the pricing page. What matters for your own sums is not our number — it is that you convert whatever number you are quoted into what this will cost my shop, with my counters and my staff, before you compare anything.
Annual billing. Most vendors, us included, discount it by roughly 20–25%. Take it only once you have run a full month on the system. A discount on something you abandon in week three is not a discount.
Line 2: Hardware — the one with the honest ceiling
Hardware is where quotes diverge most, because some vendors sell you a locked terminal and some do not.
The question to ask is not "what does your hardware cost" but "what happens when it breaks". A proprietary terminal that only runs one vendor's software means a dead device is a dead shop until that vendor replaces it. Commodity hardware means a Nairobi dealer can replace it on the same afternoon.
What a counter actually needs:
| Item | Typical range | Needed? |
|---|---|---|
| The device running the till | KSh 25,500 (handheld terminal) to KSh 45,000 (all-in-one counter PC) | Yes — unless you use a phone or laptop you already own |
| 80mm thermal receipt printer | A few thousand shillings, any ESC/POS model | Yes, for anything issuing receipts |
| USB barcode scanner | A few thousand shillings | Only if you have barcoded stock |
| Cash drawer | A few thousand shillings, printer-driven | Only if you take cash |
| Receipt rolls | Ongoing, small | Yes |
Those first two prices are sell.ke's own hardware catalogue; the peripherals are ordinary commodity items and the price depends on where you buy. The number to hold on to is that a working till does not require any of it. If your software runs in a browser and on Android, you can open on a phone or a laptop you already own, add a printer when receipts start mattering, and add a scanner when counting by hand starts hurting. Spending KSh 70,000 on day one to find out whether a POS suits your shop is the most common avoidable mistake in this list.
Line 3: Getting your catalogue in
This is the cost nobody quotes and almost everybody underestimates.
A shop with 400 products has 400 names, 400 selling prices, 400 buying prices and 400 opening stock counts to enter — 1,600 fields, assuming you know all of them, which you mostly do not. The buying prices in particular tend to live in a supplier's invoice book rather than in your head.
Three ways through it, and they cost different things:
- You do it. Free in cash, two to four evenings in time, and the arithmetic on that depends on what your evenings are worth. Realistic for a few hundred lines.
- Import a spreadsheet. Fast if you already have one that is accurate. Most "existing lists" are six months stale, which means you are doing a stock-take as well, not an import.
- Pay someone. sell.ke's done-for-you setup is KSh 10,000 and covers the catalogue, branches, staff logins, tax rates and a walkthrough. Other vendors charge for the same work, sometimes as a mandatory "onboarding fee" bundled into year one.
Whichever route, budget for it honestly. A POS with half a catalogue in it is worse than a notebook, because it reports confidently on incomplete data.
Line 4: eTIMS, and what it does and does not add
If you are VAT-registered, KRA requires your invoices to go through eTIMS. The costs attached to that depend entirely on how your POS handles it.
- If your POS submits to eTIMS directly, the cost is in your subscription tier and there is no device to buy. On sell.ke this sits on the Enterprise plan.
- If it does not, you are buying or renting a separate control unit, or typing sales into eTIMS twice, or paying a third-party bridge a monthly fee. All three are real costs and none of them appear on the POS vendor's quote.
Ask the question in this exact form: "Does the receipt my cashier prints carry the eTIMS invoice number and the QR code, generated by your system, with nothing else in the loop?" An answer that contains the word "compatible" is a no.
The deeper version of this is on our KRA eTIMS page.
Line 5: Connectivity, power and what happens without them
A cloud POS that needs a live connection to look up a price is not a till, it is a website — and Kenyan retail has outages.
The cost shows up in two places. The obvious one is the router, the backup bundle and, increasingly, an inverter or a power bank. The expensive one is the trading you lose when the queue walks because the system cannot ring up a sale.
An offline-first POS removes most of that line. Sales, price and stock lookup, stock deduction and receipt printing continue with no network and reconcile when it returns; what genuinely needs connectivity is payment authorisation, eTIMS submission and the online store. If you are comparing systems, this is the single most valuable question to test rather than ask — put the device in flight mode mid-sale and see what the screen does. We wrote up the trade-off in cloud POS vs offline POS.
Line 6: Training, and the three weeks that are slower
Every POS rollout has a dip. Your cashier was fast on the old way, is slow on the new one, and the queue notices before you do. It usually lasts one to three weeks and it costs you real money in served customers.
You can shrink it, and the levers are boring:
- Train on your own products, not on demo data. A cashier who has rung up your sukari and your unga is ready; one who has rung up "Test Product 1" is not.
- Train the awkward cases, not the easy one. A return. A discount. A split payment between M-Pesa and cash. A sale where the customer changes their mind at the last item. Those are where a cashier freezes.
- Put one person in charge. Shops that nominate a staff member as the system's owner get through the dip in days; shops where "everyone" owns it stay in it for a month.
Vendors quote training as a number of hours. The honest unit is the dip, and a vendor who tells you there is not one has not done many rollouts.
Line 7: Leaving — the cost you pay last and plan for first
Ask, before you sign: "If I leave in two years, what exactly do I get, in what format, and how long does it take?"
The answer you want is a complete export of products, stock, customers, suppliers and sales history, in CSV, available to you without asking. The answer you often get is a PDF report, or a support ticket, or silence.
This is not a hypothetical cost. A shop that cannot extract two years of sales history cannot carry its own buying patterns, its own customer list or its own margins into the next system — so it either starts from zero or it stays put and pays whatever it is asked. The exit terms are the part of a POS contract with the most leverage in them and the least attention paid to them.
The worked totals
A single shop, one counter, 400 products, already VAT-registered. Using commodity hardware and doing the catalogue yourself:
| Line | Year one |
|---|---|
| Subscription (KSh 2,999/mo, annual billing) | ~KSh 27,000 |
| Device — handheld terminal | KSh 25,500 |
| Receipt printer, drawer, scanner | ~KSh 15,000 (estimate, varies by dealer) |
| Catalogue entry | Your own evenings, or KSh 10,000 |
| eTIMS | Included at the tier that carries it |
| Connectivity | Existing bundle, plus a power bank |
| Training dip | 1–3 weeks of slower service |
| Cash out, year one | ~KSh 68,000–78,000 |
Three branches, same assumptions, with one device per branch and a shared catalogue: the subscription moves up a tier rather than multiplying by three, the hardware triples, and the catalogue work is done once. That difference — whether the second and third shop multiply your bill or not — is worth more over three years than any discount you will negotiate on month one.
What this is worth against
Costs are only half of a decision. The honest other half, stated without the usual inflation:
Stock shrinkage in an untracked shop is typically a few percent of turnover, and most of it is not dramatic theft — it is miscounts, unrecorded wastage and sales rung up wrong. A shop turning over KSh 1,000,000 a month is losing tens of thousands to that every month, and a system that makes it visible pays for itself on that line alone. Dead stock is the second one: capital sitting on a shelf that a reorder report would have stopped you buying again.
Neither is guaranteed, and you should distrust anyone who promises a figure. What is reliable is that both become visible, and you cannot fix what you cannot see. We covered the mechanics in how to reduce stock shrinkage.
The six questions to take to a quote
- Is the price per shop or per terminal, per cashier, per user?
- Is the online store included, or a second subscription?
- Does your hardware lock me in, and what happens the day it dies?
- Do you submit to eTIMS yourselves, or do I need something else in the loop?
- What exactly keeps working with no internet — show me, do not tell me?
- If I leave in two years, what do I get back and in what format?
Write down the answers. The vendor who answers all six plainly is usually the one who has thought about your shop rather than about the sale.
Want to run the numbers on your own shop? Start a free 14-day sell.ke trial — no card — or book a free demo and we will go through the arithmetic with your product list and your counters, including the lines above that are not ours.