Decisions · Tiers and status

How many tiers should your loyalty programme have

Use two published tiers for nearly every loyalty programme, and add an unpublished recognition tier only when at least one in five active members can name the behaviour that separates the top tier from the middle. A third visible tier usually increases servicing cost without changing earning behaviour, and it weakens the boundary between middle and top.

The decision

You are deciding how many published relationship bands a member can see and be sorted into, not how many benefits your programme offers. Two boundaries create one middle tier; three boundaries create two middle tiers and an extra edge case for every rule, communication and system flag. The mechanism is status contrast: a tier works only when a member can tell what behaviour separates it from the next one. With two tiers, the contrast is simple and the top tier remains scarce. With three published tiers, the middle tier tends to absorb most qualified members, and the true behavioural difference between middle and top shrinks into a service layer.

What it costs you

Running three published tiers costs more than benefit design. Each extra tier adds an eligibility rule set, a communications variant, a service script and an exception queue. The real cost is the liability of a promised tier that you cannot withdraw from members who qualified under an old rule. An illustrative calculation: suppose a middle tier earns a 100 percent points bonus and a top tier earns 150 percent. A member with 20,000 base points gets 20,000 extra points in the middle tier and 30,000 extra points in the top tier. At one cent per point, the top tier is worth 100 dollars more per member per year. A status match at 40,000 base points doubles that difference to 200 dollars per member per year. If that extra margin is not buying changed behaviour, it is paying for a status boundary that exists only in the terms.

The third tier also consumes engineering time. Every points rule, expiry rule and partner feed must be modelled for three bands instead of two. When a promotion excludes one tier by mistake, the fix goes through the same release cycle as a core redemption change. That is operating capacity diverted from benefits that members actually notice.

How to decide

Apply the one-sentence test to your current programme. Write down what a member must do to move from the second tier to the third tier. If you cannot write it without using the word more, you have a service band, not a loyalty band. Then check the gap: calculate the base points a member needs to move from the middle tier to the top tier. If that gap is smaller than the gap from base to middle, your top tier is not scarce; it is a slower on-ramp. Collapse the middle tier into the top and keep two published tiers.

For programmes with a genuine elite population, use an unpublished recognition tier instead of a published third tier. Invite only members who already show the behaviour you want, and remove the tier quietly when it stops changing behaviour. No public terms, no status match, no support queue.

What would change our mind

We would recommend three published tiers if a programme can show that its middle tier earns at least three times the points needed to enter the top tier, and that members in the middle tier still increase earning when the top tier is visible but unattainable for twenty four months. The second condition is the real one: if your membership has enough high-spending but not top-spending members who respond to a distant target, a third tier creates durable aspiration. Without that visible separation, a third tier is decoration.

Related