How hard is it to migrate?
Answer 16 questions about your legacy system and get a complexity score, a realistic duration range, team size and the migration approach that limits risk.
Same model we use in discovery
Estimated duration
Suggested team
Estimated effort
Recommended approach
Directional estimate based on Kordis Lab’s project history. Real numbers depend on a technical assessment of the code and data.
Five dimensions, one weighted score.
Each answer adds points to one of five risk dimensions. The dimensions are weighted by how often they caused delays in real migrations, then combined into a 0–100 score.
Size, technology support status, coupling and age. Obsolete stacks force a rewrite of parts you would otherwise keep.
Volume and quality. Dirty data is the most underestimated cost in any migration.
Each integration is a contract with someone else. Risk grows quickly, then flattens: 100 × (1 − e−n/10).
Documentation, automated tests and people who still understand the rules. Missing knowledge turns migration into archaeology.
Users affected, downtime tolerance and compliance obligations during the cut-over.
Base months by size (3, 6, 11, 18) × complexity factor (0.6 to 1.8) × strategy factor (rehost 0.45, replatform 0.7, refactor 1.0, rewrite 1.35). The range is −20% / +30%, because migrations overrun more often than they underrun.
Frequently asked questions
What makes a legacy migration complex?
Rarely the code alone. The biggest drivers are dirty or poorly modeled data, undocumented business rules, the number of integrations and how little downtime the business can accept during the switch.
Should I rewrite or refactor my legacy system?
Refactoring (modernizing piece by piece) is safer when the system works and the knowledge is partial. A rewrite makes sense when the stack is unsupported or the architecture blocks every change — and even then, it is safer to do it incrementally with the strangler fig pattern.
How accurate is this estimate?
It is a directional estimate, useful for budgeting conversations and comparing scenarios. A precise number requires reading the code, profiling the data and mapping integrations — which is what a technical discovery does.
You bring the problem. We figure out what needs to be built.
Tell us what’s happening in your operation, the product you want to create or the system that needs to evolve. Our first conversation starts with the problem — not the technology.
Talk to Kordis