When Not to Migrate: Fix, Buy, or Retire Instead

Good engineering leadership is often saying no to a migration. Budgets are finite. This guide helps you choose stabilize, buy, retire, or wrap—instead of a multi-year rewrite that never pays back.

Prefer stabilize when

• Incidents are rare and well-understood
• The system rarely changes and hiring is still possible
• A few high-interest debt items unlock enough speed

Invest in tests, observability, and the top 3 reliability fixes.

Prefer buy (repurchase) when

• The capability is commodity (HRIS, email, basic CRM)
• Vendors meet compliance needs at lower TCO
• Your differentiation is elsewhere

Prefer retire when

• Usage analytics show near-zero adoption
• Overlapping systems exist post-M&A
• Manual process is cheaper than keeping lights on

Retirement is a project: data archive, access removal, communication, license cancel.

Prefer wrap / facade when

• You need modern APIs for mobile/partners but core logic is stable
• Risk of touching the core is extreme short-term
• You can place an anti-corruption layer and move slowly

Migrate when

Multiple triggers from our migration triggers article stack: security risk, talent cliff, product velocity, and clear ROI within a planning horizon.

Decision worksheet

1. Cost of status quo over 3 years
2. Cost and risk of migration
3. Cost of buy/retire alternatives
4. Opportunity cost (what you will not build)

Pick the option with the best risk-adjusted return—not the most exciting stack.

Frequently asked questions

How can Auroviq help if we should not migrate?

We still help: reliability hardening, facade APIs, vendor selection, and targeted improvements that raise business outcomes without a full platform move.