Decisions · Technology

Should you modernise a legacy loyalty programme

Modernise a legacy loyalty programme only when the business can run a parallel accrual ledger for at least six months. If the existing core cannot be separated from redemption inventory, do not modernise; replace the entire programme or keep it stable.

The decision

The decision is not whether to change the member app. It is whether to separate the liability engine from the member experience before changing anything. A legacy loyalty programme usually keeps points payable, redemption rights and expiry rules inside one old system. Modernising means moving accrual and redemption onto a new ledger that the finance team can audit independently, while the old system continues to serve members until the cutover passes. The mechanism is double-entry isolation: every point issued must have a corresponding liability entry that survives the migration. If you modernise the surfaces but leave the old ledger as the source of truth, you have changed the paint and kept the crack.

What it costs you

The real cost is not software licences. It is carrying two ledgers, paying reconciliation staff and absorbing member distrust during a failed cutover. An illustrative calculation starts with a programme that accrues 36,000,000 points in a month. If the old ledger overstates balances by 2 percent, the parallel run must identify 720,000 points each month. At a redemption value of 0.5 cent per point, that is 3,600 dollars per month in audit and adjustment. Over a six month parallel run, audit cost is 21,600 dollars. Add forty engineering hours per month at 150 dollars per hour: 6,000 dollars per month, or 36,000 dollars over six months. The minimum safe cutover therefore costs 57,600 dollars before any new feature or migration tool is bought. That comes straight out of margin.

How to decide

Apply this test before any workshop. Can you freeze the legacy ledger for 48 hours without member harm? Can you run the old and new ledgers side by side and reconcile every daily batch for six months, with a named owner for discrepancies? Can you afford a full liability audit before launch, including a point by point sample of dormant accounts? If the answer to any of these is no, do not modernise. Either replace the programme with a new liability scheme or leave the legacy programme stable and reduce benefit complexity instead. A modernisation project that skips one of these tests becomes a data migration with no rollback.

What would change our mind

If the legacy ledger has already balanced to within 0.1 percent for four consecutive quarters and no redemption inventory is coupled to the old system, then a full parallel run is unnecessary and modernisation should proceed immediately. In that case the risk is low enough that freezing and auditing adds cost without reducing member harm. A second flip is regulatory: if a card network or bank requires sunsetting the old ledger by a fixed date, modernisation stops being optional and the recommendation changes from conditional parallel run to forced replacement. Then you choose the new ledger and accept a shorter reconciliation window, because the alternative is losing the ability to issue points at all.

Related