What Does It Cost to Integrate Two Systems?
A point-to-point integration between two systems typically costs SEK 100,000–500,000 to build. The price depends on the quality of the APIs, the volume of data, and how much business logic is needed between the systems. Also budget for an annual maintenance cost, since integrations need updates whenever either system changes.
Getting two systems to talk to each other sounds like a technical detail, but in practice it’s a small project of its own – with groundwork, development, testing, and a lifelong relationship with both systems’ updates. Here’s what a system integration costs in 2026 and which line items determine the price.
The price picture: SEK 100,000–500,000 for point-to-point
A point-to-point integration – a direct connection between two systems – typically costs SEK 100,000–500,000 to build. The range depends on how much needs to sync and how smart the connection needs to be.
| Integration type | Typical cost | Example |
|---|---|---|
| Simple one-way sync | SEK 100,000–200,000 | Orders transferred from e-commerce to accounting system, a handful of fields |
| Two-way sync with mapping | SEK 200,000–350,000 | Customers and inventory levels kept in sync in both directions |
| Integration with business logic | SEK 300,000–500,000 | Pricing rules, validation, error queues, and monitoring between the systems |
The groundwork you can’t skip
Most integration projects that fail don’t fail in the code – they fail in the concepts. Before anyone writes a line of code, two things need to be mapped out:
- The data models. What does “customer” mean in each system? Is an order the same thing in both? What fields exist, what format are they in, and what maps to what? The differences are always bigger than you think.
- Ownership. Which system is the master for each data type? If the same customer is changed in both systems at once – who wins? Without clear ownership rules, the integration creates duplicates and silent overwrites.
The mapping work often takes one to two weeks and feels bureaucratic, but it’s the cheapest part of the project to get right. Every ambiguity that slips through becomes a bug in production.
What drives the cost
- API quality. Well-documented, modern APIs on both ends keep the price down. Old, undocumented, or missing APIs can double the work.
- Data volume and real-time requirements. Syncing a hundred records a night is easy; tens of thousands of records in near-real-time requires queues, batching, and performance work.
- Error handling. What should happen when one system is down? Serious integrations have error queues, retries, and alerts – that’s often a third of the development work.
- Security and permissions. Personal data and business-critical information come with requirements for encryption, access control, and logging.
Maintenance: the annual budget line everyone forgets
Integrations break – not because they’re badly built, but because the systems on either end get updated. API versions get deprecated, fields change format, certificates expire, and cloud systems ship new releases without asking your integration for permission.
So budget for an ongoing line item for monitoring and maintenance. A common rule of thumb is 10–20% of the build cost per year. That sounds steep until you compare it with the alternative: an integration without an owner is often discovered only when someone wonders why orders have been missing for three weeks.
Worked example: e-commerce meets ERP
A retail company wants to sync customers, orders, and inventory levels between its e-commerce store and its ERP, in both directions. Mapping data models and ownership takes two weeks, development with error queues and alerts five weeks, testing against both systems two weeks. Total: around nine weeks and SEK 280,000–350,000, plus roughly SEK 30,000–50,000 a year for monitoring and maintenance. That’s the price tag for avoiding manual double entry – which, in this example, cost more than that in labor every year.
How to order the integration right
Four things in place before the project starts save both time and money:
- Access. API keys, accounts, and permissions in both systems – waiting for access is the most common cause of downtime.
- Test environments. Ideally a test instance of each system; testing against production data is playing with live ammunition.
- Ownership decisions. Which system is the master for each data type should be decided by the business in advance, not by the developers on the fly.
- A maintenance plan. Decide before go-live who monitors the flows and who responds when the alarm goes off.
At Weapp we build integrations as part of larger product engagements and as standalone projects. Not sure what your specific connection would cost? Get in touch with which systems are involved, and we’ll give you a range.
Frequently asked questions
Can you integrate a system that lacks an API?
Usually yes, but it gets more expensive and more fragile. The options are file-based transfer, a direct database connection, or, as a last resort, robotic automation that drives the system's interface. All three work, but they break more easily during updates – weigh the cost against upgrading to a system with a real API.
Should we build the integration ourselves or buy a ready-made connector?
If a ready-made standard connector exists between your systems, it's almost always the cheapest place to start – often a few hundred kronor a month. Custom becomes justified once you have custom fields, specific business logic, or volumes the standard connector can't handle. Start with standard and switch once it demonstrably falls short.
How long does it take to build a system integration?
A typical point-to-point integration takes 4–10 weeks from kickoff to go-live. Mapping data and ownership often takes a quarter of the time, development about half, and testing against both systems the rest. Waiting for access, accounts, and test environments is the most common cause of delays.
Why do integrations break down over time?
Because the systems on either end don't stand still. API versions get deprecated, fields change format, certificates expire, and cloud systems get updated without asking. An integration without monitoring and an owner is often discovered only after data has been missing for weeks – which is why ongoing maintenance is part of the total cost.
When do you need an integration platform instead of point-to-point?
The rule of thumb comes down to the number of flows. With one or two integrations, point-to-point is simplest and cheapest. As the number of flows grows toward five to ten, an integration platform often becomes worthwhile, since monitoring, error handling, and new connections then share the same infrastructure.