Legacy software - a system that still works but is technologically outdated - is a headache for almost every company with ten or more years of operations. The right call between upgrade, migration, and complete rewrite can mean the difference between €20,000 over six months and €200,000 over two years. Rule of thumb: if fixing the existing system costs more than 50% of new development, start from scratch. Under 30%, migration is almost always the rational choice.
Three types of legacy software
- Type 1 - works but unmodifiable. Outdated tech (PHP 5, VB6), original team gone, every change takes weeks. Approach: refactor or gradual migration.
- Type 2 - breaks more often. Bugs spread, performance drops, security updates skipped. Approach: urgent migration while stabilising the old system.
- Type 3 - no longer fits the business. Business changed; software did not. Approach: complete reassessment, often a new build.
The real cost of keeping it
“Don’t touch it” is rarely the cheapest option. Hidden costs include slower employee work (10 seconds per operation × 100 ops/day = 5 lost hours/month per employee), more errors requiring manual fixes, unpatched security holes that can cost 10× the full migration, growth ceilings, and difficulty retaining developers. This also creates GDPR compliance gaps if you handle personal data. Typical hidden annual cost for an SMB: €20,000-€80,000+.
Three migration techniques
1. Strangler Fig (gradual). New system replaces the old one feature at a time. Best for large systems where big bang is too risky. Often paired with rebuilding API integrations. 12-36 months, €40,000-€300,000+, low risk.
2. Big Bang (single cutover). Old system shuts down, new takes over. Best for smaller systems. 6-18 months development, €30,000-€150,000, high risk.
3. Parallel run. Both systems run side by side for months while users gradually migrate. Best for critical systems with zero downtime tolerance. 3-12 months parallel, €50,000-€250,000+, lowest risk.
The biggest traps in a rewrite
- Copying every old feature 1:1. Many were unused or wrong. Map actual usage first - one of the most common digitalisation mistakes we see.
- No data migration plan. Often 20-30% of total project scope, not “just a script”.
- Insufficient parallel running. Plan 2-3 months wherever possible.
- No user training. New logic without training triggers reflexive resistance.
- No post-migration plan. The Operate phase lasts years. Agree maintenance before work starts.
Frequently Asked Questions
Can we audit the legacy system ourselves? Internal teams spot functional issues but rarely deliver an objective technical assessment - incumbent developers have an emotional attachment to the system.
What if the original developers are gone? Common. The audit takes 3-5 weeks instead of 2-4. The hardest cases are obsolete stacks almost nobody uses anymore.
Can we migrate only part of the system? Yes - that is the strangler fig approach. Many teams use migration to revisit the custom vs SaaS decision per module.
How long does a typical project take? Small: 3-6 months. Mid-sized: 8-18 months. Large with deep historical data: 18-36 months.
Related Articles
- How to choose a development agency - 12 questions for migration partners.
- Custom software vs SaaS - revisit build vs buy.
- Digitalisation mistakes companies make - patterns that wreck migrations.
Have a legacy system?
Book a free Discovery call. We will review your system and propose a realistic migration plan with concrete numbers - not generic estimates.
Reach out at [email protected] or via the form on our homepage.