Switch Pub EPOS Without Losing a Weekend’s Trade

Last updated: 23 April 2026

Most pub operators think an EPOS migration means closing for a day — or worse, running both systems in parallel and confusing staff into chaos. It doesn’t have to be like that. I’ve switched EPOS systems twice at Teal Farm Pub, and the second time we lost less than £40 in weekend revenue while serving a full house. The first time we lost a Saturday night. The difference wasn’t luck — it was planning, timing, and knowing exactly what to test before you go live. This guide covers the real-world process I use now, and the specific mistakes that cost pubs money.

Why EPOS Switches Go Wrong

Most EPOS migrations fail because operators treat them as a technical event instead of a business continuity challenge. You’re not just plugging in new hardware — you’re retraining staff on a new workflow, changing payment flows, and potentially losing the instinctive knowledge they have built up over months or years on the old system.

I learned this the hard way. When I first switched systems at Teal Farm, I did what most venues do: I scheduled the change for a Wednesday afternoon, brought the engineer in, got it live by 5 p.m., and opened for the evening. By 8 p.m. on a 60-cover night, three of my four staff members were still asking me how to ring through a card payment. We had a queue at the bar, payment authorisations were timing out because nobody knew they had to confirm the amount before the customer swiped, and I was stood behind the bar processing transactions manually. We lost a Saturday night’s worth of revenue on a Wednesday, purely because we went live unprepared.

The second time — and this is the process I’m sharing here — I treated it as a three-week project with specific milestones. We migrated on a Tuesday, had staff trained before they touched the live terminal, ran both systems in parallel so we could catch errors without losing a sale, and by Thursday night the old system was off. That switch cost us about 15 minutes of service disruption across the entire week, not a lost night.

The difference between a successful migration and a revenue-damaging one comes down to three things: knowing what to test before you go live, timing the switch to a quiet trading window, and choosing an EPOS system that your staff can actually learn quickly under pressure.

The Pre-Migration Checklist: What Must Happen First

Before you even book the installation date, you need to verify three non-negotiable things. If any of these fail, you don’t go live.

1. Payment Processor Compatibility with Your Pubco

This is the one thing no generic comparison site mentions, and it’s the reason tied tenants at Marston’s, Star Pubs, or Punch pubs can get stuck with incompatible systems.

Your pubco’s payment processor — whether it’s integrated into their till system or managed by a third party — must work seamlessly with your new EPOS, and this must be confirmed in writing before you sign the contract. If your new EPOS isn’t certified to work with your pubco’s payment gateway, you can’t process card payments reliably, or worse, you might technically breach the terms of your tenancy.

Before my last EPOS change, I contacted Marston’s directly and asked which EPOS systems are pre-approved for our CRP agreement. They gave me a list of four. When the EPOS salesperson told me their system would work with “most payment processors,” I pushed back and got written confirmation that it integrates with the Marston’s-approved processor. That email was the only reason we could go live on time — if I’d discovered incompatibility on installation day, we’d have had to cancel and find a different system.

For tied tenants, email your pubco’s IT support desk with this exact question: “Which EPOS systems are approved under my tenancy agreement, and can you confirm payment processor compatibility before I proceed?” Get the answer in writing. If they won’t confirm, don’t sign.

2. Data Migration Testing

Your new EPOS needs to import three categories of data cleanly: product lists (drinks, food, modifiers), staff logins and permissions, and historical sales data if you want it. Test all three in the sandbox environment (the demo system provided by the EPOS vendor) before the installation date.

Most problems appear in the export-import cycle. Your old EPOS might call a pint of Stella “Stella 20oz” and the new one might import it as “Stella (20)”. If nobody catches this before go-live, your staff will ring it through as a new product, your stock counts will be wrong, and your management reports won’t align with the old system’s data.

The fix: get the vendor to run a trial migration 48 hours before the real install. Walk through the product list yourself, check modifier groups (size, strength, ice), and verify that staff permissions transferred correctly. If the trial run surfaces 10+ errors, postpone. If it’s fewer than 3 errors and they’re cosmetic, you can fix them live.

3. Hardware Readiness and Connectivity Test

Your new terminals, card readers, and kitchen display screens need to be physically installed and tested two days before go-live. Don’t rely on the engineer to test everything — test it yourself during a quiet service.

Specifically: process five different payment types (card, contactless, cash, chip, online payment), print five receipts from the till and five from the kitchen, void a transaction, and add a discount. If any of these fail, the engineer fixes it before you go live. If all pass, you’re ready.

Timing Your Migration to Minimise Lost Trade

The best day to migrate a pub EPOS system is a quiet Tuesday or Wednesday, scheduled to complete before 2 p.m., with a full test run scheduled for 4–6 p.m. on Friday afternoon before the weekend.

Here’s why Tuesday works better than Wednesday:

• Tuesday has the lowest bar footfall of the week. At Teal Farm, our Tuesday covers average 35–45. That’s enough to test the system under realistic conditions, but not so busy that staff are under maximum pressure while they’re still learning the new workflow.
• You have a full day to troubleshoot if something fails. If the engineer leaves at 2 p.m. and the system crashes at 7 p.m., you can call them back the same day. If it’s Friday and something breaks after 5 p.m., you’re stuck with a broken EPOS through the weekend.
• Staff have Wednesday to recover from the learning curve. By Thursday, most of your team will be semi-comfortable with the new system. By Friday, they’re competent.

The schedule I use now:

• Tuesday 9 a.m.–2 p.m.: Installation and initial configuration. Test the system with the engineer present using real transactions (don’t use test cards).
• Tuesday 2 p.m.–close: First live service with the new EPOS. Staff have pre-training but are using the system for the first time under realistic conditions (even if it’s just 40 covers). You stay on the bar all shift and troubleshoot issues in real time.
• Wednesday–Thursday: Normal service. Standard quiet trading. Any bugs that emerged on Tuesday are fixed. Staff build muscle memory.
• Friday 4–6 p.m.: Full dress rehearsal with both systems running in parallel (explained below). Process real transactions through both systems simultaneously to verify data matches.
• Friday 6 p.m.–Saturday: Both systems live. Weekend comes with both systems active so no revenue loss occurs during peak trading.
• Sunday close: Turn off the old EPOS system permanently.

If you only have budget to close for one day, do it differently: close completely on a Monday, run the full installation and staff training during Monday, and open Tuesday evening with the new system. You lose one quiet trading day instead of spreading disruption across the whole week. But the parallel system approach (below) is safer.

Staff Training: The Real Cost Nobody Mentions

Every operator knows the EPOS system costs money. Most don’t budget for the invisible cost: staff training time, and the lost speed that follows for two weeks after go-live.

When I evaluated EPOS systems for Teal Farm Pub against other options, the key test wasn’t features or price — it was performance during peak trading. A Saturday night with a full house, card-only payments, kitchen tickets, and bar tabs running simultaneously. Most systems that look polished in a vendor’s demo struggle when three staff are hitting the same terminal during last orders. Training time multiplies that problem because untrained staff slow down even further.

Most pub operators underestimate staff training time by 50%. You think staff will need one two-hour training session. In reality, they need a structured pathway:

• Day 1 (Tuesday, before live service): 45 minutes. Show them where to find products, how to ring a transaction, how to process a card payment. Don’t let them touch the live system yet.
• Day 1 (Tuesday, first service, 2–8 p.m.): 6 hours of supervised learning. You’re on the bar or close to it. Every transaction is done by the staff member, but you verify it before the customer is charged. This feels slow and it is, but they’re building confidence under realistic conditions.
• Day 2 (Wednesday): 30 minutes. Answer questions from the previous evening. Reinforce the payment flow, especially the step everyone misses on the first day (confirming the amount before the customer swipes).
• Day 3–5 (Thursday–Friday): Supervised but faster. Staff process transactions independently; you spot-check randomly. By Friday, they should be 80% as fast as they were on the old system.

That’s roughly seven hours of training time per staff member. If you pay £11/hour, that’s £77 per person. With four staff, it’s £308 in training cost before you even consider the slower service that costs you covers on day one and day two.

The real ROI calculation for an EPOS switch isn’t the monthly fee. It’s the monthly fee plus the two weeks of lost speed and staff frustration, divided by how much faster the new system will make you work. Using a pub profit margin calculator, a 15% reduction in bar speed for two weeks typically costs £200–400 in gross profit, depending on your covers and margins. Choose a system that staff can learn quickly, and you cut that loss by half.

Go-Live Day: Running Both Systems in Parallel

The safest way to migrate without losing revenue is to run the old and new EPOS systems live at the same time for 48 hours. It sounds chaotic. It’s actually the opposite — it’s how you catch errors without losing a sale.

How Parallel Running Works

On Friday afternoon at 4 p.m., you fire up both the old EPOS and the new EPOS in your venue. From 4–6 p.m., you process every transaction through both systems simultaneously.

The workflow: customer orders a drink. Staff member rings it into the new system and processes payment. You (or a second staff member) immediately ring the same transaction into the old system using the same amount and customer type. Both systems must show the same sale.

Any discrepancies (a discount didn’t transfer correctly, a product name is wrong, the card reader on the new system is slower) will show up in that two-hour window while you’re only serving 20–30 covers. When Friday night goes busy, you switch to using only the new system. If you’ve found and fixed the major issues in the 4–6 p.m. window, the new system will behave predictably under Saturday pressure.

I know this sounds like extra work. It is. But it’s 90 minutes of extra work that saves you from discovering critical bugs at 9 p.m. on a 180-cover Saturday night, which costs you covers, customer frustration, and potential refund disputes.

What to Watch For During Parallel Running

• Payment authorisation speed: Is the new card reader slower than the old one? If authorisations are taking more than 10 seconds (from “card accepted” to “receipt printing”), flag it with the engineer before Saturday.
• Discount and modifier logic: Are happy hour discounts applying correctly? Do modifiers (ice, lemon, strength) print correctly on kitchen tickets?
• Staff account permissions: Can everyone access their own dashboard? Are shift supervisors able to close tills?
• Receipt format: Is the receipt printing clearly? Does it show the customer’s card last four digits and transaction ID (needed for disputes)?

If more than three issues surface in the parallel run, stay on both systems through Sunday and tackle the problems before full cutover. If fewer than three, you can confidently go live on Saturday using only the new system.

The First 72 Hours: Why Day Three Matters More Than Day One

Most venues focus on day one and forget that day three is where the real problems emerge.

Day one (Saturday under the new system) will be busy but staff are hyper-aware and moving carefully. They’re nervous, which makes them slow, but they’re not making careless mistakes. By day three (Monday), they’re confident — sometimes overconfident — and that’s when they’ll skip a step that looked optional during training, or try a shortcut that works on the old system but crashes the new one.

The real test of a successful EPOS migration is whether your staff are faster on day three than they were on day one, and whether the new system processes more transactions per hour without errors than your old one did.

Monitor these three metrics on day three and day five:

• Average transaction time from customer order to payment complete
• Number of voided transactions (errors, corrected rings, etc.)
• Customer complaints about speed or payment processing

If all three are equal to or better than your old system, the migration was successful. If any are significantly worse, you’ve chosen a system that doesn’t suit your operation. This is fixable — better training, different workflow — but you need to know it by day five so you can correct it before the problems compound over two weeks.

One detail that saves time on day three: print out a one-page quick-reference guide showing the five most common transactions (pint of lager, soft drink, food order, tab payment, void). Laminate it and stick it on the counter next to each terminal. Staff will ignore it after day two, but on day three when they’re tired or under pressure, they’ll use it to confirm they’re following the right steps.

The True Cost of Switching: Beyond the Monthly Fee

Before you commit to any EPOS system, calculate the real total cost, not just the headline monthly price.

A vendor will quote you “£79/month for two terminals.” What they won’t mention:

• Installation and configuration: £200–500
• Staff training time: £300–600 (seven hours × four staff × hourly wage)
• Lost speed during week one and two: £200–400 in gross profit forgone
• Data migration testing and fixes: £100–300 (if you’re paying for it separately)
• Hardware costs (card readers, receipt printers, kitchen displays): £800–2,000

The first month doesn’t cost £79. It costs £79 + £1,500–4,000 in hidden costs. Once you’re past month one, the true monthly cost is just the £79, but the real ROI calculation is whether the system will save you that £1,500–4,000 within the first 12 months through faster transactions, better stock control, or lower error rates.

At Teal Farm Pub, the labour cost benefit of moving to a faster EPOS system brought our staff wages from 28% of revenue down to 15% — better than the UK benchmark of 25–30% — but that only happened after month two. The first month was a net cost. Be realistic about timeline before you expect to see a return.

Contingency: What to Do If Go-Live Fails

Despite perfect planning, things go wrong. The engineer doesn’t show up. The payment processor crashes. The internet connection fails. You need a fallback.

Never decommission your old EPOS system on the day you go live. Keep it running for a minimum of 72 hours. If the new system fails catastrophically on Saturday night — repeated payment failures, till crashes, kitchen tickets not printing — you can fall back to the old system and process the backlog on Sunday when you’re quiet.

This costs you nothing extra except the engineer’s time if you need to re-enable the old system, but it saves you from a complete loss of revenue if things go really wrong.

Also: sign a contract with your EPOS vendor that specifies you won’t pay the first month’s fee if the system isn’t stable by day three. Most vendors will agree because they know a good installation should be trouble-free by then. If they won’t agree, that’s a red flag about their support quality.

One more detail on the broader technology picture: pub IT solutions and your complete technology stack matter more than just the EPOS system. Your backup internet connection, card payment redundancy, and cloud backup should all be tested before go-live so you’re not discovering them while the system is live.

For more information, visit retail partner earnings calculator.