Decisions · Strategy and governance

Should you align every function before launching loyalty

No. Launch once the functions that control point liability and member recognition are aligned, then align the rest after launch. Waiting for every function costs more in lost margin and data than the rework it prevents.

The decision

The decision is not whether alignment matters, but how much of it to buy before go-live. Alignment is a queue. Marketing needs member identity rules, finance needs accrual liability logic, operations needs redemption handling, and customer service needs adjustment authority. Each function you sequence before launch adds a dependency. The mechanism is a trade: pre-launch alignment buys lower rework and lower liability, but it pays for that with delayed earning and delayed member data. The question is where to stop the queue.

What it costs you

The recommended path accepts post-launch cost in three places: liability from misapplied points, engineering time to retrofit unaligned functions, and member trust that erodes when service is inconsistent. Those costs are real, but they are usually smaller than the opportunity cost of waiting. An illustrative calculation: suppose a programme issues points with a redemption value of one cent each. Customer service is not aligned at launch, and a batch error credits 240,000 points incorrectly. The point liability is 240,000 x 0.01 dollars = 2,400 dollars. Retrofitting customer service later takes one engineer for two months at 12,000 dollars per month, which is 24,000 dollars. Total cost of the recommended path is 2,400 plus 24,000 = 26,400 dollars. Delaying launch for four months to align customer service before go-live would forgo 7,500 dollars gross margin per month, or 30,000 dollars. The post-launch path is cheaper by 3,600 dollars.

How to decide

Apply a manual workaround test to each function. List the functions that touch point issuance, redemption, liability accounting, and member support. For each one, ask: can this function run safely for three months with a documented manual process? If yes, leave it for post-launch alignment. If no, align it before launch. A function fails the test if a single manual error can create unrecoverable liability or a legal breach, such as expiry logic or redemption value. A function passes if the workaround is simply slower, such as campaign analytics or partner reporting. Launch when no remaining unaligned function can create liability greater than the gross margin lost by a one-month delay.

What would change our mind

We would reverse this recommendation if a signed partner contract pays a launch bonus only when all named functions are aligned, and that bonus is at least double the cost of the additional alignment work. Under that condition, each marginal function added before launch yields a direct, measurable return rather than a speculative reduction in rework. The queue then should run to the end, because the last function aligned still contributes more than it costs. If the bonus is absent or smaller than the retrofit cost, the original recommendation stands.

Related