How to choose a POS for a small shop

Most POS comparisons are feature checklists, and a feature checklist is the least useful way to choose one. Every product on your shortlist will claim billing, stock and reports. What separates them is what happens on a bad day — and a bad day in a shop is not an unusual event.

Here are the questions that actually decide whether you will still be happy in two years.

1. What happens when the internet goes down?

Ask this first, and do not accept "it syncs later" without a demonstration. There are three honest answers and they are very different:

• It stops. The till is a web page talking to a server. No line, no sale. Common, and rarely stated plainly.
• It queues. Sales are held on the device and pushed later. Better — but stock is not checked while it is queued, so two counters can sell the same last packet, and you find out at reconciliation.
• It does not need the internet at all. The database is on the computer in your shop. The connection is for updates and licensing, not for selling.

The third is the only one where a week-long outage is a non-event. It is also the one that costs the vendor the most to build, which is why it is rarer than the marketing suggests.

2. Where does your data physically live, and how do you get it out?

If the answer is "our cloud", ask two follow-ups. Can you export everything — sales, products, customers, purchases — not just a summary? And what happens to it if you stop paying?

A great many shops discover the answer to the second question at the worst possible moment. Software you have stopped paying for should still leave you holding your own history. Ask for a one-click full backup you can take today, before you have anything to lose.

3. What is the real cost over three years?

Compare the total, not the monthly figure, and include:

• The per-till or per-user charge, multiplied by the counters you will have in two years, not the one you have now.
• Whether reports are in your tier. Some products put the reports you need for tax behind a higher plan.
• Whether you keep access to your own history if you stop paying.
• What an extra device costs the day you add a second counter.

A subscription is not automatically worse than a one-off price. What is worse is a subscription you must keep paying to read what happened last year.

4. Can it be wrong, and would you know?

This is the question almost nobody asks, and it is the one that costs real money. Three specifics:

Can two tills sell the same last item? In a shop with one counter this never comes up. In a shop with two it comes up every busy Saturday. The answer should be that the software makes it impossible, not that it is unlikely.

Can a cashier change a price? If yes, is it recorded, and can you cap it? "Charging less" is where margin quietly leaves a shop, and it should always be somebody's recorded decision rather than an invisible one.

Does last month's profit change when this month's costs change? If the software recalculates margin from today's cost price, a supplier price rise silently rewrites history you already looked at. Each bill should remember what its goods cost on the day.

5. Does it fit YOUR trade?

A pharmacy needs a controlled-drug register and expiry dates. A kiryana store needs weight and credit. A clothing shop needs one product in twelve sizes. A general product will do all of them adequately and none of them well.

Look for two things: whether the trade-specific parts genuinely exist (ask to see them, not to hear about them), and whether the parts you will never use can be switched off. A pharmacy field on a clothing form is a small annoyance a hundred times a day.

6. What does support actually mean?

Get specific. Is there a person, a phone number, and a working day? What happens if the computer dies on a Friday — can you move your licence yourself, or do you wait for a human? Can you restore a backup without help?

Software for a shop with no IT department has to be recoverable by the shop.

A short practical test

Whatever you shortlist, do this before you buy:

1. Install it and put twenty of your own real products in, with your real prices.
2. Ring up ten sales, including one on credit and one return.
3. Pull the network cable and try to do it again.
4. Take a backup, and open the file to confirm something actually came out.
5. Close the day and count the drawer.

An afternoon of that tells you more than a month of feature comparisons. If a vendor will not let you do it, that is itself the answer.

SalezFlow POS is built for the third answer to question one — it runs on your shop's own computer and does not need the internet to sell. If you want to run the test above, download it and try it against your own shop, or tell us what you sell and we will tell you straight whether it fits.