Swiss mid-market ERP programmes rarely collapse. They drift. The go-live moves by a quarter, then another, and the business case quietly stops being mentioned. By the time anyone calls it a failure, the money is already spent.
Across the engagements we review, the technology is almost never the binding constraint. The constraint is decision latency: how long it takes an organisation to answer a question the implementation cannot proceed without.
That matters more now than it did five years ago, because many programmes starting in Switzerland today have a date imposed from outside. Vendor support for widely used systems ends in 2027 and 2028, and a fixed date makes slow decisions more expensive, not less.
Where the time actually goes
A stalled programme looks busy. Standups happen, tickets move, consultants bill. What is not moving is the set of open questions that block configuration.
| Phase | Planned | Typical actual | Main driver of the gap |
|---|---|---|---|
| Blueprint | 8 weeks | 14 weeks | Process ownership unclear |
| Configuration | 16 weeks | 22 weeks | Decisions reopened after sign-off |
| Data migration | 6 weeks | 13 weeks | Source data quality discovered late |
| UAT | 4 weeks | 9 weeks | Test cases written after the build |
The figures are typical of the programmes we have assessed, not a benchmark, and the pattern matters more than the numbers. Nobody owns the answer, so the answer arrives late, and everything downstream absorbs the delay. The data migration line is usually the largest relative overrun, for reasons set out in ERP data migration.
Most of these programmes are replacements rather than first installations. Across the EU, 57.6% of manufacturing enterprises with ten or more employees already used ERP software in 2025, according to Eurostat. A replacement carries the old system's data and the old system's habits, and every habit is a question someone has to answer.
What is decision latency?
Decision latency is the elapsed time between a question being raised by the implementation team and a binding answer being given by the business. It is the single most predictive number we track, and almost nobody measures it.
Measure it for two weeks. Log every blocking question, who it went to, and when it came back. The distribution will tell you more about your go-live date than the project plan will. A log needs five columns and no tool: the question, the date it was raised, the person who owns the answer, the date of the answer, and whether the answer was later reopened. The last column is the one that surprises people.
Why are so many Swiss companies starting ERP programmes now?
Because vendor support is ending for systems many of them run. SAP has said mainstream maintenance for the core applications of SAP Business Suite 7 ends at the end of 2027, with optional extended maintenance to the end of 2030 at a higher fee. Microsoft's extended support for Dynamics NAV 2018 ends in January 2028.
The SAP position is set out in its 2020 maintenance announcement: mainstream maintenance to the end of 2027, then optional extended maintenance from 2028 to the end of 2030 at a premium of two percentage points on the existing maintenance basis, and a commitment to maintain SAP S/4HANA to the end of 2040. Microsoft's lifecycle page for Dynamics NAV 2018 shows mainstream support ended in January 2023 and extended support ends in January
- In our experience both are widely installed in Swiss companies of 50 to 500
people.
A deadline changes the behaviour of a programme in a predictable way. The blueprint is compressed to protect the date, so the hard process questions are deferred into configuration, where answering them costs rework. The deadline is a reason to start the decision work earlier, not to skip it. Where extended maintenance is available, its premium is a known number to set against the cost of a rushed go-live, and that comparison belongs in front of the owners before the plan is signed.
Three fixes that move the date
- Name a single decision owner per process area. Not a committee. One person who can say no on Friday and have it stick on Monday.
- Put a clock on open decisions. A question older than five working days is escalated automatically, not discussed at the next steering committee.
- Freeze the scope by decision, not by date. A scope freeze that does not close the underlying questions just moves the argument later.
None of this requires new software. It requires someone with the authority to close questions and the discipline to do it on a cadence. In a 300-person company the owner for order-to-cash is usually the head of sales operations or the CFO's deputy; in an 80-person company it is often the managing director, which is fine as long as the clock still applies to them.
How long should a mid-market ERP programme take?
For a Swiss company of 100 to 500 people with one primary legal entity, a realistic first go-live is nine to fourteen months from blueprint in the programmes we have assessed. Programmes that promise six months usually exclude data migration, and programmes planned for two years usually absorb decision latency into the schedule rather than fixing it.
Below 100 people, on a product built for smaller companies, four to eight months is more typical, with the same caveat about data. Either way, plan the go-live against the Swiss calendar as well as the project plan. SIX publishes a new release of the Swiss Payment Standards every February, which banks support from November, so a November go-live tests two changes at once. Year-end closing, the VAT period boundary and the busy season are the other dates to avoid.
What does a stalled ERP programme cost?
Every month of slip costs the implementation partner's run rate, the old system's licence and support, and the time of key users who are tied up but not finishing. In the programmes we have assessed, the partner's run rate is the largest visible line and the key users' time the largest invisible one, because nobody invoices for it.
Work out the monthly figure before the programme starts, not when it is already late. Partner days per month times the day rate, plus the old system's monthly running cost, plus the hours of your key users valued at what replacing them would cost. For a company of 150 people this number is usually large enough to change how fast the steering committee is willing to decide. It is also the number that makes the case for a decision owner with real authority, as the ROI framework for digital transformation explains.
The old system does not disappear at go-live either. Accounting records must stay readable for ten years under the Code of Obligations, Art. 958f, and the Federal Tax Administration expects the VAT audit trail from voucher to return to remain traceable. A programme that slips past the old system's support date keeps paying for both.
Should we customise or adopt the standard process?
Adopt the standard unless the process is genuinely a source of competitive advantage. The test is commercial, not technical: if a customer would not notice or pay for the difference, it is not worth the maintenance cost of a customisation across every future upgrade.
The cost is more frequent than it used to be. Microsoft's Business Central, in its online version, receives two major updates a year, in April and October, plus minor updates in the other months, and every customisation has to keep working through each of them. On SAP, the maintenance commitment to 2040 means a customisation built now may be carried for well over a decade. Each one is also a decision someone has to make, document and defend, which is decision latency by another route.
The Swiss questions only the business can answer
Some decisions recur in almost every Swiss programme, and none of them can be delegated to the implementation partner. Each is small. Together they are dozens of questions, and they cost days if owners are named before the blueprint and weeks if they surface during configuration.
Languages. Which languages customer documents, article texts and screens run in, and who maintains each translation. A manufacturer selling into Romandie and Ticino needs French and Italian documents; someone has to decide whether every description is kept in every language or only the ones that get printed.
Addresses and payments. The QR-bill has accepted only structured addresses since November 2025. Someone has to decide who splits the old free-text addresses in the customer and supplier masters, and by when.
Tax codes. Mapping old VAT codes to new ones, including history from before the 2024 rate change, is a finance decision with an audit trail behind it, not a configuration detail.
Entities and payroll. With an EU subsidiary, which processes are harmonised and which stay local. And whether Swiss payroll runs inside the ERP or in a separate product on Swissdec's list of certified payroll software, which changes the integration scope.
The data side of these decisions is covered in more detail in ERP data migration.
Does this apply to a company of 40 people?
More sharply, not less. A small company has fewer people who can answer a blocking question, and those people are also running the business, so a question waits behind a customer call rather than behind a committee. The mechanism is identical; the queue is shorter and slower.
The fix scales down cleanly: one named owner, a five-day clock, and questions closed in writing. No governance framework required. If you are not sure a company of that size needs an ERP yet, the signals are in when a small company actually needs an ERP.
What to do before the programme starts
Decide who can say no, and tell the implementation partner. Most contracts name a project sponsor and a steering committee, and neither is the same as a person who can close a process question inside a week. Writing that name down before kick-off is the cheapest schedule protection available.
If you are still choosing the system, the selection method that reduces later decision load is in ERP selection for the Swiss mid-market. If you are choosing who helps you run it, ask how they handle a decision that is not being made, a question covered in how to choose a management consulting firm in Zurich, and check rates against what a Swiss company should pay for consulting.
Where to start this quarter
Pick one process area. Instrument decision latency for a fortnight. Bring the distribution to your steering committee instead of a status colour. The conversation changes immediately, because the number is not arguable.
For an industrial company, order-to-cash or plan-to-produce is usually the right first area, because it touches the most people and the most open questions. Our ERP consulting practice instruments this in the first fortnight of an engagement. If the programme is part of a wider change, the first quarter of that is sketched in digital transformation for an SME: the first 90 days.
Sources
- Eurostat — E-business integration, Statistics Explained, ERP use by economic activity, 2025
- SAP — SAP S/4HANA maintenance until 2040, and the choice for SAP Business Suite 7, maintenance dates and extended-maintenance premium
- Microsoft — Dynamics NAV 2018 lifecycle, mainstream and extended support dates
- Microsoft — Business Central update cycles, two major updates a year
- SIX — ISO 20022 and the Swiss Payment Standards, annual release cycle and transition deadlines
- SIX — QR-bill, structured addresses since November 2025
- Swissdec — Certified ERP and payroll software, register of certified products
- Fedlex — Swiss Code of Obligations, SR 220, Art. 958f: ten-year retention
- Federal Tax Administration (ESTV) — MWST-Kontrolle, the VAT audit trail
Related reading
- ERP selection for the Swiss mid-market
- ERP data migration: the part that overruns
- When does a small company actually need an ERP?
- How to measure ROI on a digital transformation before you fund it
- ERP consulting
Working on this right now?
Tell us where you are in two questions. A consulting partner reviews every enquiry within 2 business days.
