If nine out of ten of your loyalty members receive exactly the same treatment, you do not have a loyalty programme; you have a deferred-discount scheme with a database attached.
India is saturated with such schemes. Every airline, hotel chain, fashion retailer, fuel brand, and increasingly every fast-moving consumer goods (FMCG) company runs one, and the mechanics have converged to the point of interchangeability: points on spend, tiers on volume, redemption against the same inventory the member would have bought anyway. Meanwhile the subscription programmes of the large e-commerce and food-delivery platforms have reset consumer expectations of what membership should feel like — immediate, tangible, worth paying for — which makes the points-and-tiers formula look even thinner. Yet Indian boards keep funding these programmes, usually justified by retention claims nobody has measured. The justification that would actually hold — that in an indirect-distribution economy the loyalty programme is often the brand’s only mechanism for capturing consented, identified, transaction-linked consumer data — rarely appears in the business case at all.
Across a decade of deployments, the pattern repeats, and it was already documented when we were doing customer-experience strategy work for luxury hospitality in the mid-2010s. The industry research circulating then made uncomfortable reading, and its shape has not materially improved; treat the figures as illustrative of the era’s survey findings rather than as current audited data.
| Loyalty programme failure signal (industry surveys, mid-2010s, illustrative) | Approximate finding |
|---|---|
| Members at risk of switching brands | ~30% |
| Member spend going to brands other than their “preferred” one | ~50% |
| Customer relationship management (CRM) strategies judged unsuccessful by their own managers | ~60% |
| Companies believing their loyalty programme lacks differentiation | ~90% |
The last line is the killer fact, and it has two readings. The comfortable reading is that programme design needs refreshing. The correct reading is that when ninety percent of companies concede their programmes are undifferentiated, the reward mechanics cannot be where the value lives — identical mechanics cannot differentiate anybody. What remains is the data: a loyalty programme is a standing consent to be known. It attaches an identity to transactions in categories where, especially in India’s intermediary-dominated distribution, no other touchpoint does. Judged as a marketing promotion, the typical programme is a cost centre accruing a points liability on purchases that would have happened anyway. Judged as a first-party data capture vehicle feeding segmentation, personalisation, and media targeting, it can be the highest-yielding asset the brand owns. The same programme, valued two ways — and most Indian operators are running the first valuation on an asset only the second valuation justifies.
The evidence for the second valuation is old and consistent. The clearest use case in loyalty analytics is targeted winback: identify lapsed or at-risk members from their behavioural record, target them with a specifically relevant offer, and measure recaptured revenue against programme cost. Published vendor case studies from the era reported such programmes returning several times their cost — figures best treated as directional orders of magnitude, but the mechanism is sound and we have seen it reproduce: the margin between a blanket offer and a behaviourally targeted one is where the entire economic case for loyalty analytics sits. None of it is available to a programme that treats its member file as a mailing list.
The deeper discipline is matching treatment to the member’s actual state of intent. In the hospitality work, the operative segmentation was four traveller mindstates: the guest who knows exactly which hotel they want; the one who knows the destination but not the hotel; the one who wants a break but has no destination; and the one not thinking about travel at all. Each mindstate calls for a different investment — conversion friction removal for the first, competitive presence for the second, inspiration content for the third, patient brand-building for the fourth. The most common failure in the category, then and now, is applying mindstate-one mechanics to mindstate-four members: firing direct-booking prompts and redemption reminders at someone who is not planning anything, training them to ignore the brand entirely. An undifferentiated programme commits this failure by construction, at every send, to ninety percent of its file. The brand lens says loyalty should feel like recognition; being sent an offer that plainly ignores everything the brand knows about you is the opposite of recognition. The evidence lens says the same thing in numbers: response rates on undifferentiated loyalty communication decay toward the floor, and the decay is usually logged in the email platform nobody reviews.
There is one honest India-specific caveat. In hospitality specifically, loyalty economics are structurally harder here — corporate travel booked through travel managers and online travel agencies (OTAs) weakens the individual member’s decision power, and acute price sensitivity means tier perks rarely beat a cheaper rate. That caveat sharpens the thesis rather than blunting it: if the reward mechanics face structural headwinds, the data asset is the only part of the programme guaranteed to hold value. And in indirect categories — FMCG, spirits, pharmaceuticals-adjacent wellness — where the brand never sees the consumer transaction at all, the loyalty programme is not one capture channel among several; it is frequently the only one, which is precisely why it must be designed for data yield rather than borrowed from an airline template.
On Monday, run two audits. First, the differentiation audit: how many behaviourally distinct member segments received materially different treatment — different offer, different content, different channel, different cadence — in the last quarter? If the honest answer is one or two, your programme is the ninety percent, and every point issued is liability without return. Second, the asset audit: restate the programme’s value in data terms — consented, contactable, transaction-linked profiles; their reachability; the incremental behaviour attributable to targeted treatment versus blanket offers, measured with holdouts. Then decide like an investor. If the data yield justifies the points liability, fund the analytics layer that the member file has been waiting for. If it does not, redesign the programme around capture and differentiated treatment — or shut it down and stop paying for a discount scheme wearing a loyalty badge.