Ask what loyalty marketing returns against acquisition marketing and you will get a number. Ask three sources and you will get three numbers, none of which agree, all of which are quoted with equal confidence. The 2026 crop includes loyalty technology investment returns cited around 4.8x, first-year program ROI ranges of 8x to 22x, and a target net program ROI of 5x or better.

Those are not competing estimates of the same quantity. They are estimates of different quantities that have been flattened into a single sentence by repetition, and a marketing team that budgets against the wrong one will either overinvest badly or dismiss the category entirely. It is worth separating them before using any of them.

What Each Number Is Measuring

“Technology investments in loyalty deliver 4.8x returns” is a statement about platform spend — the software, the integration, the data infrastructure. The denominator is the technology line, not the program. It says nothing about the cost of the rewards themselves, which in most programs is the larger number by a wide margin.

“First-year ROI of 8x to 22x on implementation costs of $15,000 to $40,000” is a first-year figure on a small implementation base. First-year returns on a new program are structurally inflated because the initial enrolment wave captures guests who were already returning. Those guests were going to book again; the program is now taking credit for them. This is the single most common overstatement in loyalty measurement and it is why year-two numbers routinely disappoint against year-one.

“Target net program ROI of 5x or better” is the most useful of the three, because net program ROI subtracts program costs from incremental revenue. The word doing the work is incremental — revenue that would not have occurred without the program. That is the only version of the number that answers the budgeting question.

None of these is a like-for-like comparison with acquisition spend, and any headline claiming a clean multiple between the two should be treated with suspicion.

The Comparison That Does Hold Up

Where the retention case is genuinely strong is on the cost side, and it is strong enough not to need inflation.

Customer acquisition costs in hospitality approach roughly $50 per booking. Against that, a well-performing property sees around 25% to 30% of its business from returning guests, while loyalty members contribute approximately 35% of occupancy. Put plainly: about a third of the rooms are being filled by guests whose marginal acquisition cost is close to zero.

Loyalty members also spend more once they are there — research puts the range at 12% to 25% more annually than non-members across hospitality settings. That is a real, measurable, and reasonably well-attested effect, though it carries its own selection problem: guests who join loyalty programs were probably higher-frequency guests to begin with.

The most actionable data point in the current set is a reallocation one. Properties are reported to have doubled marketing ROI by shifting 20% of acquisition budget into loyalty. That is a marginal-reallocation claim rather than a total-return claim, and marginal claims are much more useful for budgeting than headline multiples, because budgeting decisions are marginal decisions. Nobody is choosing between all-acquisition and all-retention. They are choosing what to do with the next 20%.

The Measurement Trap

The reason loyalty ROI reporting is so unreliable is that the default attribution is wrong in a specific and predictable way.

A returning guest who books through the loyalty channel gets attributed to the loyalty program. Some proportion of those guests would have returned regardless — they liked the property, it is near their office, it is the one they always use. The program did not cause that booking; it captured it. Attributing the full revenue to the program inflates the return by however large that proportion is, and in mature programs it is large.

The correction is to measure incrementality rather than attribution, which requires a holdout: a matched group of eligible guests who do not receive the program treatment, compared against those who do. Very few hospitality programs run this, because it feels like deliberately not marketing to good customers. It is the only method that produces a number a CFO should act on.

A workable middle path for properties unwilling to run true holdouts is to measure at the margin instead — track what changes when a specific campaign runs against a specific segment, rather than trying to value the program as a whole. Campaign-level marginal measurement is imperfect but it is honest, and it is directly actionable.

What This Means for the Budget

Three practical conclusions follow.

Do not fund a loyalty program on a headline multiple. Fund it on your own acquisition cost per booking and your own repeat rate, both of which you already have. If acquisition is running near $50 a booking and repeat business is under 20% of occupancy, the reallocation case makes itself without needing an industry statistic.

Separate the technology decision from the rewards decision. They have different returns, different risk profiles, and different failure modes. A property can get most of the recognition and first-party data benefit — the parts the 2026 trend commentary keeps pointing at, including instant recognition, personalization, and first-party data strategy — with a modest technology spend and a restrained rewards structure. The expensive failure is the reverse: a generous rewards liability on infrastructure that cannot identify who is earning it.

Budget for year two. The year-one number will be good and will overstate the case. A program evaluated on its first-year return and then cut in year three when the number normalizes has cost more than never starting, because the rewards liability outlives the enthusiasm.

The underlying case for retention spend in hospitality is sound and does not depend on any particular multiple being true. It rests on a cost asymmetry that any property can verify from its own books in an afternoon: it costs real money to find a new guest and almost nothing to keep one who already likes you. Everything past that is measurement discipline.