How to revive a failing loyalty program

Programs die quietly because the headline numbers — cumulative registrations, lifetime scans — only ever go up. The honest signals are trends: 30-day actives falling for three consecutive months; redemption dropping below ~50% of earned value (points nobody redeems are points nobody values); activity concentrating into a shrinking set of counters while whole territories go silent; and the most telling one, which never appears on a dashboard — the field force stops mentioning the program because it no longer helps them sell. Any two of these together mean the program is dying, whatever the cumulative chart says. The full metric ladder and healthy ranges are in our KPI guide; this article is about what to do when the ladder is broken.

Data lens (week 1–2). Cohort the entire base: never-activated, lapsed (no scan in 90 days), fading (declining frequency), healthy. Cut each cohort by geography, supplying dealer, tier and enrolment vintage. Reconstruct the redemption funnel — earned → eligible → attempted → fulfilled — and find where value stalls. Reconcile the points liability: unredeemed balances are both a finance exposure and, handled well, the fuel for the relaunch.

Field lens (week 2–3). Structured interviews with 30–50 participants across cohorts — not a survey, conversations: "When did you stop scanning? What happened?" The answers are usually brutally specific: the payout that never came in March, the catalogue mixer that took two months, the new salesman who never mentioned the program. Interview the field force and dealers separately; ask the salesmen what they actually say at counters, and listen for the program's absence.

Money lens (week 3–4). Where did the budget actually go — how much reached active participants vs leaked to fraud, breakage-by-neglect and administration? Audit the fraud history honestly: some "failing" programs actually died of over-zealous fraud controls that blocked legitimate users, and some died because gaming by a few big counters poisoned the economics for everyone. Check TDS hygiene too: unfiled 194R liabilities on old payouts surface at the worst possible time — during a relaunch.

The audit ends with a one-page verdict: which of the five killers (usually two or three in combination), which cohorts are recoverable, and what the fix costs against the budget bands in our budgeting guide.

Fix quietly when mechanics are sound and the failure is operational: payout speed, catalogue quality, communication cadence. Ship the fixes, then market the improvement inside the existing brand — "ab payment turant" is a stronger message than a new logo.

Relaunch under a new name when trust itself is broken — when the audit interviews use words like "chalu" or "fraud" about the old program. Incremental fixes inherit stigma; a rebrand resets the conversation. But a rebrand only works under two conditions. First, the old program's obligations are honoured before the new name appears: balances migrated or paid, disputes settled. A relaunch that starts by confiscating history is the old program with better graphics, and the trade will say so. Second, the relaunch is visibly different in the one dimension that killed the original — if payouts were slow, the new program's signature must be money-in-seconds demonstrated live at every relaunch meet; if the app was the problem, the new program should need no app at all.

Migrating balances, step by step. Freeze the old scheme on an announced date. Audit balances and cap clear outliers pending investigation (legacy books hide old gaming). Publish each user's opening balance personally over WhatsApp with a 2–3 week dispute window, and be generous on small disputes — fighting over ₹500 costs more than it saves. Credit accepted balances into the new program, or better, pay a portion out in UPI immediately: opening the relaunch with real money landing in accounts is the single most persuasive act available. Model the tax first: paid-out legacy balances plus restart bonuses count as 194R benefits, and can push many users past the ₹20,000 per-PAN threshold in one quarter — run the exposure through the TDS calculator before announcing amounts.

Segment before spending. High-value lapsed counters (the top 10–15% by historical volume) get personal field visits — their objections are specific, fixable, and worth an RSM's afternoon; a counter that left over a stuck payout returns when someone brings proof it is fixed and settles the old amount on the spot. The broad middle gets a WhatsApp win-back journey: acknowledgement ("aap ne scan karna band kar diya — hum samajhte hain kyun"), the fix, a low-friction restart bonus (2x points on the first five scans, or a ₹50 bonus on the first scan back). Never-activated users are not lapsed — they were never onboarded; they get the full first-time journey, assisted where possible.

Two field truths about win-back. The most convincing message is not an offer but proof from a peer — a payout screenshot from the counter next door outperforms every bonus structure; seed each market's revival through its most-followed counters and let WhatsApp do the rest. And protect the relaunch from the old fraud patterns: lapsed-user bonuses are a fresh gaming surface (mass re-activation of mule accounts is a known pattern), so velocity caps, device checks and UPI-name matching apply from day one of the win-back, not after the first incident.

Communicating the relaunch. Sequence exactly as a launch: dealers first (with an honest accounting of what went wrong and what their role in the new design is), then the field force with retrained scripts, then counters and influencers through meets where the first live scan-to-UPI happens on stage. Set expectations for the first 90 days publicly — payout SLA, helpdesk number, what happens to old balances — and then run the relaunch with the same war-room discipline as a new program's hypercare. A revived program gets one second chance from the trade; there is no third.