Insight: legacy migration

What legacy migration really costs, and why big programmes overrun.

Legacy migrations have a reputation for blowing budgets. The cause is rarely the technology. It is the shape of the work. Here is what drives the price, what waiting costs, and how to take the risk out.

Published 7 minute read

Bars of rising height with a gold trend line, representing the cost of legacy systems over time

Why do legacy migration programmes overrun?

Legacy migration programmes overrun mainly because of their size. A multi-year scope hides unknown dependencies and poor-quality data until late in the programme. Time-and-materials billing turns each surprise into additional cost. And a single big-bang cutover at the end concentrates the risk into one weekend that nobody wants to sign off. Smaller, fixed-price steps expose problems early, while they are still cheap to fix.

Australia has its own cautionary tale. The Queensland Health payroll system went live in March 2010 with 2,422 known defects, according to the Queensland Government, which later said the system was set to cost taxpayers in excess of A$1.2 billion. Most organisations will never face a failure on that scale, but the pattern behind it is common: a large scope, an unclear baseline, and a cutover the business was not ready for.

What does waiting cost?

Doing nothing feels free. It is not.

  • Support premiums. SAP's optional extended maintenance for ECC after 2027 costs an extra two percentage points on the maintenance basis. Microsoft's Extended Security Updates for products such as SQL Server 2016 are a paid add-on.
  • Security exposure. Once security updates stop, every newly found vulnerability stays open.
  • Specialist skills. Skills in COBOL, RPG, VB6 and older SAP releases are increasingly scarce as experienced people retire, and rates reflect it.
  • Blocked projects. New channels, analytics and AI stall when the data they need is locked in a system nobody wants to touch.

The United States Government Accountability Office reports that federal agencies typically spend about 80 per cent of their IT budgets on operating and maintaining existing IT, including legacy systems. In 2019, GAO identified ten critical legacy systems that, according to the agencies, ranged from about 8 to 51 years old and together cost about US$337 million a year to operate and maintain.

What drives the price of a migration?

Cost driverWhy it mattersHow to reduce it
IntegrationsEvery interface must be found, moved and testedInventory them first; retire the unused; wrap the rest behind APIs
Data volume and qualityBad data fails loads and reconciliations lateProfile early; archive what you do not need; fix at the source
Custom codeEach customisation must be understood, moved or replacedRetire unused code before moving the rest
TestingUntested migrations fail in productionAgree acceptance tests and reconciliation rules up front
Cutover constraintsTight windows need more rehearsalMeasure downtime in rehearsals, not estimates
Scope changesUncontrolled change is the biggest overrun driverFix the scope per step, and move new requests to the next step

How fixed-price sprints change the economics

A fixed price for a defined scope moves the overrun risk from you to the supplier. Splitting the work into 30-day steps adds three more protections:

  1. Problems surface in weeks, not years. Bad data and hidden dependencies appear in the first sprint, while they are cheap to fix.
  2. Every step is live and signed off. Value lands monthly, and the business case can be checked against real results.
  3. You can stop between sprints. If priorities change, you keep everything already delivered and owe nothing for sprints you have not started.
The cheapest migration is not the one with the lowest quote. It is the one whose price is fixed before the work starts.

Questions to ask any migration supplier

  • Is the price fixed for a written scope, and what happens if you take longer?
  • What will be live, and signed off, in the first 30 days?
  • How will you prove the data is complete and correct?
  • What is the rollback plan, and has it been rehearsed?
  • Do we own all code, data, documentation and credentials at the end?
  • Are you paid more if we choose a particular product or licence?

How we help

Sovereign Systems Labs delivers legacy system migration as fixed-price 30-day sprints from A$50,000, after a five-day Migration Blueprint from A$9,500 that is credited in full if you go ahead within 60 days. Prices are in Australian dollars and exclude GST. Each sprint runs old and new in parallel, reconciles the data with your owners, and cuts over with a rehearsed rollback.

Sources

FAQ

Quick answers

Anything else? Email contact@sovereignsystemslabs.com.

How much does a legacy system migration cost?

It depends on the number of integrations, the volume and quality of data, how much custom code there is, testing needs and cutover constraints. Sovereign Systems Labs prices each defined scope as a fixed-price 30-day sprint from A$50,000, confirmed in writing after a five-day Migration Blueprint.

Why do legacy migration projects go over budget?

Mostly because of their size. Large scopes hide unknown dependencies and poor data until late, time-and-materials billing turns each surprise into extra cost, and a single big-bang cutover concentrates risk at the end.

Is it cheaper to keep the old system?

In the short term, often. Over time, extended support premiums, paid security updates, specialist skills and the projects the old system blocks usually cost more than a well-run migration.

Book a free call