Legacy core modernization in waves, not slogans

The graveyard of core replacements is full of programs that were ‘90% done’ for two years. The last 10% is products with odd interest logic, dormant accounts, and the batch that only two people understand.

Wave by product and channel. Keep the old core posting until recon is boring. Give operations a veto on the next wave. Budget the dual-run as a first-class cost, not a surprise.

Partners who promise a 18-month full replacement without naming which products stay behind should be asked to show the last time they did it.