Choosing a POS System: Hardware, Payments, and the Lock-In Question

POS systems are sold on features — inventory, loyalty, staff scheduling, gorgeous reporting dashboards — and paid for somewhere else entirely. For most businesses, the payment processing bundled into the system costs several times more per year than the software subscription, and it's the part nobody compares because it arrives as a percentage rather than a price. Add proprietary hardware that becomes scrap if you leave and data that's awkward to export, and a decision made on a thirty-minute demo becomes a five-year commitment. This guide reverses the order: processing economics first, lock-in second, features third — which is how the total cost actually stacks up.

The processing question comes first

Before a demo, before a feature comparison, ask one question: does this system require me to use its payment processing, or can I choose my own?

Closed processing means the POS and the merchant account are inseparable. The rate you're quoted is the rate you have, and your only leverage is leaving the entire system — which the hardware and data situation is designed to make painful. Open processing means the software works with independent providers, so you can shop the rate, negotiate at renewal, and change processors without changing systems.

Why this dominates everything else: at meaningful card volume, a modest rate difference produces annual costs that dwarf the software subscription. A business processing steadily through the year can easily pay more in the difference between a competitive and an uncompetitive rate than it pays for the POS itself — and that difference recurs every year, growing with the business. The fee anatomy in our processing costs guide explains what's actually in that percentage; the point here is simply that a bundled rate you can never shop is a permanent tax on growth.

None of this means bundled processing is always wrong. Integrated systems can be genuinely simpler, reconciliation is cleaner, and support is single-threaded. But the tradeoff should be made knowingly, at a rate you compared — not discovered two years later when you try to negotiate and learn there's no alternative to negotiate with.

Total cost of ownership, calculated properly

Build the comparison on a spreadsheet with your real numbers, not the vendor's example business. Five components:

Component	What to calculate
Processing	Annual card volume × effective rate, plus per-transaction fees × transaction count. Use your actual average ticket — small tickets are dominated by fixed per-transaction fees, large tickets by percentage.
Software	Monthly subscription × 12, times the number of registers or locations, plus any per-module add-ons for features you actually need.
Hardware	Upfront purchase or lease payments, plus a realistic replacement cycle. Leased hardware over a multi-year term frequently exceeds purchase cost.
Implementation	Setup, data migration, menu or catalog build, training, and the productivity dip during the first weeks.
Exit	Early termination fees, remaining lease obligations, hardware you can't reuse, and the migration cost to move again.

Two mistakes to avoid in the math. Don't compare monthly subscription prices in isolation — a system that's cheaper monthly and more expensive on processing is more expensive, full stop, and vendors know which number gets quoted in the ad. And model your average ticket honestly: flat-rate pricing that looks excellent on a $60 restaurant check looks very different on a $9 coffee, where fixed per-transaction fees consume a large share of the sale.

Hardware: owned, rented, or trapped

Three arrangements, with very different consequences at renewal.

Standard hardware — off-the-shelf tablets, common receipt printers, generic cash drawers — is the most flexible position. If you change software, most of the equipment still works, and replacements are available from ordinary suppliers at ordinary prices.

Proprietary hardware is purpose-built for one vendor's platform and typically functions only with it. It's often excellent equipment, and it is also a switching cost disguised as a purchase: leaving means re-equipping every station. When evaluating, ask directly whether the terminals work with any other software, and price the answer into the exit column of your total cost model.

Leased hardware deserves particular scrutiny. Equipment leases in the payments industry are frequently long, non-cancelable, and separate from the POS agreement itself — meaning you can leave the software and still owe payments on the terminals for years. Read whether the lease is with the POS vendor or a third-party finance company, and check the total payments against the outright purchase price; the multiple is often surprising.

Worth knowing as a baseline: a great deal of card acceptance no longer needs dedicated hardware at all. Phones can now accept contactless payments directly, as our tap-to-pay guide covers, which makes an expensive terminal fleet a genuine choice rather than a requirement — and gives you a fallback if a terminal fails mid-service.

Features that actually matter by industry

Feature lists are long because they're written for everyone. The features that matter are the ones touching your daily workflow, and they differ sharply by business type.

• Restaurants and bars: table and course management, modifiers and kitchen routing, split checks and shared items, tip handling and pooling, and — increasingly decisive — how well online ordering and delivery integrations work without a tablet farm behind the counter.
• Retail: real inventory management with variants, purchase orders and receiving, barcode workflows, multi-location stock visibility, and clean e-commerce synchronization so the shelf and the website agree.
• Service businesses: appointment scheduling, staff calendars and commissions, client history, deposits on booking, and the ability to take payment remotely as well as in person — the virtual terminal and payment link capabilities that let you capture a deposit on a phone call.
• Everyone: offline mode that still takes payments when the internet drops, reporting you can export, permissions granular enough to limit staff access, and a support model that answers during your trading hours rather than during theirs.

Two evaluation disciplines that consistently pay off. Run your three most awkward transactions through a trial — the split check with a coupon, the partial return with an exchange, the order modified after it's fired — because that's where systems reveal their design assumptions. And ask your staff, not just the owner: the person using it two hundred times a day will identify friction the demo never surfaces, and staff resistance is a common reason implementations fail.

The contract terms to negotiate

The features are marketing; the contract is the actual relationship. Read these before signing, and negotiate them — most are more flexible than they appear, particularly for a business bringing real volume.

1. Term length and auto-renewal. Multi-year terms that renew automatically unless cancelled in a narrow window are standard and worth pushing back on. At minimum, calendar the cancellation window the day you sign.
2. Early termination fees. Know the number. A fee that scales with remaining months is the mechanism that makes an unsatisfactory relationship expensive to leave.
3. Rate change rights. Can the provider change processing rates during the term, and with what notice? Pass-through increases from networks are normal; discretionary margin increases should not be.
4. What happens to hardware if you cancel — returned, kept, or still being paid for.
5. Service levels and support hours, especially whether support exists at the hours you actually trade.
6. Data ownership and export, in writing. This is the sleeper term, covered next.
7. Reserve and funding terms if processing is bundled — the holds and reserve mechanics in our funding guide apply here exactly as they do to any merchant account, and they belong in the diligence.

Your data and whether you can take it

Over years of operation, a POS accumulates the most valuable dataset a small business owns: customer records and contact details, complete sales history, inventory and cost data, and staff performance. If that data can't be exported in a usable format, you're locked in by your own information regardless of what the contract says about termination.

Ask specifically, before signing: Can I export customer records, sales history, and inventory data on demand, in a standard format like CSV, without paying a fee or filing a request? A vendor confident in its product answers yes immediately. Hesitation is informative.

Two adjacent points. Stored payment credentials are a separate question from POS data — if you keep cards on file for repeat customers, whether those tokens are portable determines whether a switch means asking every customer to re-enter their card, the migration issue our orchestration guide treats as the central architectural decision. And run a real export periodically rather than trusting the feature exists — an export you've never tested is a promise, not a capability.

Switching without closing for a day

Most POS migrations go badly for predictable reasons, and the fixes are logistical rather than technical.

1. Export and clean your data first. Migration is the natural moment to remove discontinued items, deduplicate customers, and correct pricing — but do the cleaning before the import, not after.
2. Pick a genuinely slow period. Never a holiday season, never a launch week. The productivity dip is real and lasts days.
3. Build and test the full catalog or menu before cutover, including modifiers, taxes, and discounts. Ring up your twenty most common transactions on the new system before it goes live.
4. Train before, not during. Staff learning a new system in front of a queue produces errors, slow service, and resentment that outlasts the migration.
5. Run parallel briefly if you can — old system available as a fallback for the first days, particularly for payment acceptance.
6. Verify the money path end to end on day one: a real transaction, a real refund, a real settlement into the bank account, and a matching report. Reconciliation problems discovered a month later are painful to unwind.
7. Keep access to the old system's reporting for at least a year, for taxes, disputes, and historical comparison.

Frequently asked questions

---

This guide is educational. POS features, pricing structures, and contract terms vary widely by vendor and change frequently; review current agreements carefully before signing.