A yoga studio owner asked me last week which loyalty model to run. She had been told by three different consultants and gotten three different answers. Here is the actual decision tree, with worked Filipino café and salon examples for each model, and the hybrid that works when none of the three pure models fit.
Punches
What it is: visit-based. Show up 10 times, get a reward. Counted by visit, not by spend.
Best for: cafés, salons, brow bars, pet groomers, juice bars, nail studios, anywhere the average ticket is small and visits are the main unit of value. Roughly 60% of small Filipino service businesses we work with land here.
Why it works: the unit (a visit) matches what the customer is buying (a visit). Stamping is fast. Progress is visible. The math is obvious to anyone, owner and customer.
Where it fails: businesses with wildly different basket sizes. A retail shop where one customer spends ₱200 and another spends ₱2,000 shouldn't reward them with the same stamp.
The numbers to set:
- Stamps to redeem: typically 8–10 for cafés, 5–6 for salons, 4–5 for higher-ticket services.
- Reward value: aim for the reward to cost you 8–12% of the cumulative spend. A 10-stamp coffee card at ₱150 average means ₱1,350 of paid revenue, so a ₱150 free drink is 11% (about right).
- Min spend per stamp: set a floor (e.g. ₱100 minimum) so a customer can't accumulate stamps by buying water bottles.
Points
What it is: spend-based. ₱1 spent earns 1 point. Earn enough points, redeem for something.
Best for: retail shops, restaurants with a wide menu, hardware stores, pharmacies, anywhere the basket varies a lot. Roughly 25% of the shops we work with land here.
Why it works: it scales with value. The customer who spends three times as much earns three times the rewards. Big spenders get rewarded for being big spenders.
Where it fails: small-ticket businesses. A coffee shop running a points system makes the customer wait 60 visits for a free drink. The math is correct; the feel is wrong. Customers can't easily convert "350 points" into "what does that get me." They lose interest.
The numbers to set:
- Earn rate: 1 point per ₱1 spent is the easiest to explain. Anything else (1 per ₱5, 1 per ₱10) requires the customer to do mental math, which they won't.
- Reward thresholds: a small reward at 250 pts (~₱250 spend), medium at 800 pts, hero at 1500 pts. Multiple thresholds mean the customer always has a "next milestone" to look at.
- Welcome bonus: 50–100 pts at sign-up. Not enough to redeem for free, but enough to put them visibly closer to the first reward.
Tiers
What it is: status-based. Cross a threshold of spend or visits, move into a higher tier with better benefits.
Best for: spas, fitness studios, premium services, club-style businesses (wine bars, members-only gyms), anywhere the customer wants to be treated as a regular and where there's a real difference between casual visitors and devoted members.
Why it works: it triggers status motivation. People who would never chase a 5% discount will chase Gold-tier access. The motivation is identity ("I'm a Gold member here") more than value.
Where it fails: walk-in heavy businesses. Tiers need repeat visits to be felt. A first-timer doesn't care about silver-vs-gold. If 70% of your traffic is one-and-done, tiers are wasted on most of it.
The numbers to set:
- 3 tiers, never more. Bronze → Silver → Gold. People can't track 5. They can track 3.
- Thresholds based on annualized spend or visits. Aim so 60% of members hit Bronze (the bottom), 30% hit Silver, 8% hit Gold, 2% are at-rebuild risk. The pyramid matters because it makes Gold feel scarce.
- Real differentiation between tiers. Bronze gets points + birthday. Silver gets free express service every quarter. Gold gets a private appointment line + an annual gift. If the only difference is "10% off vs 15% off," nobody chases the upgrade.
The 30-second decision tree
- Is your average ticket under ₱300 and your customers come weekly? Punches.
- Is your average ticket above ₱500 and varies by 3× or more between baskets? Points.
- Are 50%+ of your revenue from members visiting 6+ times a year, and there's a "VIP" feel to the higher-ticket tiers of service? Tiers.
- None of the above clearly fits? Default to punches. They're the easiest to explain, the cheapest to run, and the easiest to migrate away from later if your data tells you to switch.
Three worked examples
Example 1: a Cebu specialty café (Punches)
Average ticket: ₱180. Visit frequency for regulars: 2.5×/week. 9-buy, 10th-free punch card to a free signature drink (₱220 menu price). Per-customer cycle:
- 9 paid drinks × ₱180 = ₱1,620 revenue
- Free 10th drink: COGS ₱45 (beans, milk, cup)
- Net cycle: ₱1,575 revenue with ₱45 reward cost = 2.7% effective discount
- Customer reaches reward in ~25 days, completes the program 14 times/year
Yearly impact per regular: ₱22,000 in revenue with ₱630 in reward cost. The card pays for itself in week 1.
Example 2: a Manila hardware store (Points)
Average ticket: ₱850. Basket spread: ₱100 (one screw) to ₱5,000 (full bathroom set). 1 point per ₱1, three reward tiers:
| Reward | Threshold | What it costs them | What it costs you |
|---|---|---|---|
| ₱100 store credit | 500 pts | ~3 visits | ₱60 (60% margin) |
| ₱300 store credit | 2,000 pts | ~10 visits | ₱180 |
| Free delivery (1 trip) | 3,500 pts | ~18 visits | ₱200 |
Why points fits here: the customer who walks in for ₱100 still earns proportionally to a customer who walks in for ₱5,000. A punch card would over-reward small baskets and under-reward big ones.
Example 3: a BGC spa (Tiers)
Average ticket: ₱1,800 per visit. Visit frequency varies wildly (some 1×/year, some 1×/month). 3-tier program based on annualized spend:
| Tier | Threshold | Perks | % of base |
|---|---|---|---|
| Bronze (default) | 0–₱15,000/yr | Birthday treat, 1 pt per ₱1 | 62% |
| Silver | ₱15–35K/yr | Above + free express manicure quarterly + early booking on weekends | 30% |
| Gold | ₱35K+/yr | Above + dedicated stylist + VIP appointment line + ₱2,000 anniversary gift | 8% |
The Silver-to-Gold gap is intentionally costly. Once a customer hits Silver and sees what Gold gets, the upgrade pulls hard. Their cumulative spend increases by 35% in the year they're chasing Gold (real number from one we work with).
When to combine models
About 15% of the shops we work with end up running a hybrid. The most common one is punches + tiers: every customer is on a punch card, and tier status sits on top giving extra perks. Earned by visit count, not spend.
This works for businesses where:
- Visits are the main unit (so punches are right) but
- You want a way to reward your most frequent regulars with status, not just a free drink (so tiers add the differentiation)
Example: a Cebu nail salon. Standard 5-stamp punch card (every 5 visits = free express manicure). Plus 3 tiers based on visits/year: Bronze (1–9 visits), Silver (10–24), Gold (25+). Bronze gets the punch card and birthday. Silver gets the punch + a free callus treatment quarterly. Gold gets the punch + private appointment line + free polish change between full services.
The risk with hybrids: complexity. If you can't explain the program in two sentences, your customers can't either. Most owners should run a single model first for at least 90 days, then layer the second only if the data clearly says they need it.
When to change your model
Don't change inside the first 90 days. The signals that say it's time to switch:
- Redemption rate stays under 10% after 90 days. Either your reward is wrong or your model is wrong. Try fixing the reward first (cheaper), then the model.
- Average ticket distribution skews wider than 3:1. If your top quartile of receipts averages 3× the bottom quartile, punches are over-rewarding small baskets. Time to consider points.
- You can't differentiate VIPs from casuals. If your top 10% of spenders feel they're getting the same program as everyone else, tiers fix that. Layer them on top of your existing model, don't replace.
What most small Filipino businesses end up with
Most of the shops we work with start with punches and stay there. It's not because punches are best in theory. It's because they're easiest to explain ("buy 9, get 10"), easiest to use (the cashier just taps "Award stamp"), and they map cleanly to how customers actually think about coming back ("I've got 7, just need 3 more").
If you can't decide between two models, pick punches. You can layer points or tiers later if your data shows you need them. The cost of starting with punches and switching is small. The cost of starting with a complex tier system that nobody understands is months of dead air while your customers wonder if they're members of anything at all.
This week
- Pull your last 100 receipts. Calculate the median ticket and the spread (top quartile vs bottom quartile).
- Match against the decision tree above. Pick one model.
- Set your reward + threshold per the section's worked example.
- Run it for 90 days without changes. Then audit.