Xero Integration in New Zealand

Check the App Store before you build anything

Xero has a large marketplace, and for a lot of common connections — Shopify, WooCommerce, several job management and inventory systems — something already exists that costs $30 a month instead of $15,000.

A developer worth hiring will tell you this in the first conversation, and will lose the job cheerfully if an off-the-shelf connector does what you need. Ask every one of them: is there an app that already does this? If nobody mentions the App Store, be curious about why.

Custom integration is for the cases the marketplace does not cover: your own software, an unusual process, or a connector that does 80% and leaves the awkward 20% to a person.

What you are really buying

People assume the cost is in talking to Xero. It is not — the API is good, well documented, and a competent developer can push an invoice into it on the first afternoon.

The cost is in your business rules, and there are always more than anyone expects:

• Which account code does each product or service map to?
• How is GST treated on freight, on discounts, on international sales?
• What happens when a customer exists in Xero twice, spelled slightly differently?
• What does the system do with a part-paid invoice, or a credit note against an invoice that has already been reconciled?
• Which system wins when both have been edited?

Every one of those is a decision only you and your accountant can make, and every one has to be encoded. That is the project. The developer who asks you the most uncomfortable questions during quoting is usually the one who has done this before.

The question that separates the quotes

What happens when Xero is unavailable?

It will be. Xero has maintenance windows, occasional outages, and rate limits that a busy system will hit. What the integration does in that moment is the difference between an integration you forget about and one that quietly loses invoices.

A good answer sounds like: the work is queued, retried with backoff, and if it still fails it appears somewhere a human will see it, with a button to retry that cannot create a duplicate. A bad answer is that it will not happen.

Ask about duplicates specifically. The most common failure in a rushed integration is the same invoice arriving in Xero twice after a retry, and finding that at month end is genuinely painful.

One direction is much cheaper than two

Pushing invoices into Xero is well-trodden and predictable. Keeping records in step in both directions is a different class of problem: both systems can change the same contact, and something has to decide which change wins.

Before paying for two-way sync, check whether you actually need it. Often one system is genuinely the master for each kind of record — your ordering system owns products, Xero owns financial data — and one-way flows in each direction are far simpler than true synchronisation.

What good looks like after launch

You should be able to answer three questions without ringing your developer:

1. Did today's invoices make it into Xero?
2. If something failed, what and why?
3. How do I retry it safely?

If the integration has no screen that answers those, you have bought something that works until it does not, and then becomes a mystery. Ask to see what that screen looks like on a project they have already delivered.

Comparing five quotes

The prices will cluster more than in most categories, because the work is well understood. Read for judgement instead:

• Who asked about your account codes and GST treatment before quoting?
• Who mentioned the App Store?
• Who described failure handling without being asked?
• Who has done a Xero integration in your industry before?

Describe the two systems you want connected and what is being typed in twice today. Developers who work with the Xero API regularly will tell you whether it is a $10,000 job, a $40,000 job, or a $30-a-month app you did not know existed.