Choosing a POS system in Kenya: seven questions to ask first

Every point-of-sale vendor will show you a smooth demo on a fast laptop with clean data. That tells you almost nothing about how the system behaves on a Friday evening with a queue at the till and the internet down.

Here is what we suggest you ask instead.

1. What happens when the internet drops?

Ask the vendor to unplug the network mid-demo. A system built for Kenyan retail keeps selling offline and syncs when the line returns. One that freezes is one that will cost you a day's takings the first time Safaricom has an outage.

2. Who can void a sale, and is it logged?

Shrinkage rarely looks like theft. It looks like voided sales, discounts applied after the fact and items removed before the receipt prints. Every one of those actions should be tied to a named user and a timestamp you can report on.

3. Can I see yesterday's numbers from my phone?

If checking your own takings requires being physically at the shop, you will stop checking. That is when problems grow.

4. How does M-Pesa actually reconcile?

• Does the system match the till number payment to the specific sale?
• Or does someone type the code in by hand afterwards?

The difference is hours per week and a reconciliation you can trust.

5. What does month-end look like?

Ask to see a stock valuation, a slow-moving report and a gross-margin-by-category report. If these do not exist, you are buying a cash register, not a system.

6. Who owns the data?

You should be able to export everything to Excel whenever you want, without asking permission. Be careful with any vendor who cannot demonstrate that in thirty seconds.

7. What happens when I need a change?

Every business outgrows its first configuration. Find out now what a change request costs and how long it takes, because you will be making one within six months.

A demo shows you the best day. Ask about the worst one.

If you would like us to walk you through these against your own products and prices, book a demo and we will load a sample of your data first.