5 tier design mistakes that break program economics
Tier design isn't "let's call the levels Silver/Gold/Platinum and pick some thresholds". It's financial engineering that determines whether the program pays back in 12 months — or drains resources for three years.
Across 10+ loyalty implementations in hospitality, we've seen five repeating failures. All are fatal to program economics.
Mistake #1: Thresholds below the average guest
"To attract as many people as possible, let's set Silver at 5 nights". Sounds logical. In practice — most guests already do 5 nights a year without any program. You're giving them status and benefits for free.
Right approach: Silver threshold sits above "average guest" behaviour. The goal is to push +30% of guests to extend stays or increase frequency in order to reach the tier. If 60% of guests qualify for Silver "by virtue of last year's behaviour" — the program just gave money away without changing behaviour.
How to calculate
Look at nights-per-guest distribution for the past year. Take median, add ~30%. That's your Silver threshold. Gold — median + 80%. Platinum — top 10%. Elite — top 2%.
Mistake #2: "White elephant" tiers with no economic meaning
"Big chains have 4-5 tiers, let's do the same". But if the difference between tiers is "bottled water in the room" vs "double amenity set" — that's not a tier, that's ambition inflation.
Right approach: every tier must have economically distinct value to the guest. Welcome — first-time benefits. Silver — +X% earning, basic upgrades. Gold — guaranteed upgrades, F&B credit. Platinum — suite preference, dedicated curator. Elite — event invitations, partner deals.
A tier is valid if you can answer: "What behaviour does this tier push/reward?" If the answer is unclear — the tier is redundant.
Mistake #3: Spend-only qualification
"Tier = how much you spent". Simple. Problem — the hotel pushes guests to spend more per visit, but not more often and not for longer.
In hospitality you need 4 qualification axes (Monetary / Volume / Frequency / Engagement). A guest who bought a 1-night $1,500 stay and a guest who spent 30 nights for $1,500 are wildly different guests with different LTV. The program must distinguish.
Right approach: formulate thresholds as "OR-rule": reach Silver either by spending X, OR by doing Y nights, OR by Z stays. This lets the program reward different patterns: business guest with short frequent visits AND a leisure family with one big annual trip.
Mistake #4: "Everyone loses status every year"
Strict annual reset is an airline practice. Works poorly in hotels.
Hospitality cycles are longer: a guest may skip a year for personal reasons (childbirth, relocation, illness). Losing Platinum after 4 years of loyalty over one missed year is guaranteed churn.
Right approach: grace period 6-12 months + tier protection after 3 years at a level. More complex to model — but retention of a 4-year Platinum × their LTV more than pays it back.
Mistake #5: No per-tier ROI calculation
The most critical. A tier is defined without answering: "How much will we pay to keep a guest in Gold, and how much will they bring?"
Concrete example: Gold offers +20% earning + guaranteed upgrade + F&B credit. Cost of these benefits averages $180/guest/year. If average Gold member generates $2,300/year revenue vs $1,500 Silver — incremental $800. Cost $180. ROI 4.4×.
Run this calculation for every tier. Somewhere ROI comes out 5×, somewhere 2×, somewhere 0.8× — that means the tier needs to be either eliminated or its benefits redesigned.
What we do in the discovery phase
Over 2 weeks of discovery we build tier design based on:
- Distribution of nights/spend/stays of your guests over 24 months (pulled from PMS)
- P&L modeling of benefit cost per tier
- Sensitivity analysis: what happens to economics if thresholds shift ±20%
- Defending the model in front of CFO with numbers, not narrative
This is not what comes with loyalty-software vendor onboarding. It's methodological work — without which software turns into an expensive Excel.
