A retail or hospitality business commonly runs a POS at the counter, a booking platform online, a payment terminal, and an accounting package behind all of it. Each does its job. None of them knows what the others know, so a person becomes the integration.
01 What disconnection actually costs
The obvious cost is retyping. The expensive costs are quieter: takings reconciled days late, a customer who exists three times under slightly different spellings, stock figures nobody trusts, and end-of-month taking a full day because the numbers have to be assembled before they can be checked.
None of that appears as a line item, which is why it survives for years.
02 Pick a source of truth first
Before any building, decide which system owns each kind of record. Customers usually belong to the booking platform or the CRM. Sales belong to the POS. Financial records belong to the accounting package.
Skip this and you get a sync loop: the POS updates a customer, that flows to the booking system, which flows back, and an old address quietly overwrites a new one. It is the single most common failure in integration work and it is a decision, not a technical problem.
Data should flow one way for each field unless you have a specific reason otherwise. Two-way sync is occasionally necessary and always more expensive to build and to reason about. Start one-way and add the reverse only where the business genuinely needs it.
03 Option 1: native integrations
Check this before paying anyone for anything. Most established POS and booking platforms already publish to Xero or MYOB, and payment terminals like Tyro and Square integrate with common POS systems out of the box.
- Cost: usually included, sometimes a small monthly add-on.
- Good for: daily sales totals into accounting, payments matched to sales.
- Limits: you get the fields the vendor decided on. If your workflow needs something they did not anticipate, you are stuck.
04 Option 2: an automation platform
Where nothing native exists, a platform like Zapier, Make or n8n sits between the systems and moves records as things happen. A booking is made, and the customer appears in the CRM and the mailing list without anyone typing.
- Cost: a subscription that scales with volume, plus the build.
- Good for: most small and mid-sized businesses, most of the time.
- Limits: it depends on each vendor's API. If a platform has no API and no webhooks, this route does not exist.
Choosing between those three platforms is its own decision, and the pricing model matters more than the feature list at high volume.
05 Option 3: a direct build
Worth it when volumes are high enough that per-task pricing hurts, when the logic is genuinely specific to your business, or when a system will only talk in a way the off-the-shelf platforms cannot handle. It costs more up front and nothing per transaction.
We took this route for Chargeroo, where payments, venue systems and customer records had to behave as one platform rather than four connected ones.
06 Traps worth knowing about
- Silent failure. A broken sync usually does not announce itself. Alerting matters more than the integration.
- Duplicates. Decide how you match records — email, phone, or an ID — before the first sync, not after.
- Historical data. A new integration handles new records. Existing mess needs a separate one-off migration, and it is worth budgeting for.
- Timing. Real-time is not always right. Nightly batches are simpler, cheaper and perfectly adequate for accounting.
- API changes. Vendors change things. Something that has run for a year will eventually break.
If you already know which systems need connecting, the audit is short. See systems integration for what we cover, or tell us what you are running and we will say which of the three routes fits.