Cost by integration type
| Integration type | Typical cost | Timeline |
|---|---|---|
| Single well-documented REST API, read-only | £2,000 – £6,000 | 1–2 weeks |
| Two-way sync with a modern SaaS API | £6,000 – £18,000 | 3–6 weeks |
| Accounting platform (Xero, Sage, QuickBooks) | £8,000 – £25,000 | 4–8 weeks |
| Payment or banking integration with reconciliation | £12,000 – £35,000 | 6–12 weeks |
| ERP integration (SAP, NetSuite, Dynamics) | £25,000 – £90,000 | 3–6 months |
| Legacy SOAP, flat file, or screen-scraped system | £15,000 – £60,000+ | 2–5 months |
The spread within each row is mostly about data volume and error tolerance. Syncing two hundred contacts nightly is a different job from syncing two hundred thousand orders in near real time where a duplicate creates a double charge.
What actually drives the price
Documentation quality comes first. A well-documented API with an OpenAPI spec, a sandbox, and clear error codes can be integrated in a fraction of the time of one where the documentation is a PDF from 2019 and the only way to learn the behaviour is to call it and see what happens.
Authentication is the next factor. A static API key is trivial. OAuth 2.0 with refresh token rotation, per-tenant consent, and a token store that survives a deploy is real engineering. Some enterprise systems require certificate-based authentication or a VPN tunnel, which brings infrastructure and security review into scope.
Then there is the data model gap. Your "customer" has one email address; theirs has a contact hierarchy with roles. Your order statuses are five values; theirs are twenty-three, four of which are deprecated but still returned. Mapping that is where integration projects actually spend their time, and it cannot be estimated accurately until someone has seen real data from both sides.
The costs quotes usually leave out
Line items that turn a fixed price into a change request:
Reconciliation and monitoring are the two most commonly omitted and the two that matter most in production. An integration nobody is watching is an integration that has probably been broken for a while.
When to use an integration platform instead
Tools such as Zapier, Make, and Workato are genuinely the right answer for low-volume workflows between popular SaaS products where the logic is simple and a few minutes of latency is fine. Setup is hours rather than weeks, and a non-developer can maintain it.
They stop being the right answer when volume grows, when you need transactional guarantees, when the transformation logic gets complex enough to be unreadable in a visual builder, or when per-task pricing overtakes the cost of building properly. A workflow running a hundred thousand tasks a month on a per-task plan will usually cost more per year than a custom integration costs once.
A reasonable pattern is to prototype on an integration platform to prove the workflow and learn the data, then rebuild the parts that turn out to be high volume or business critical.
Ongoing cost
Budget 15 to 25 per cent of the original build cost per year for maintenance on any integration you depend on. APIs change, credentials expire, vendors deprecate endpoints with three months notice, and volumes grow past assumptions made at design time. Integrations are not a one-off purchase; they are a small permanent commitment.
AyTech note: The safest projects start with a narrow, measurable workflow, then expand after real users prove the value. This keeps budgets controlled and gives Google, buyers, and stakeholders clearer proof of expertise.
Need a practical technical plan?
AyTech can review your requirements, map the risks, and turn the idea into a scoped delivery plan.
Custom software services