Design a Spend Threshold Without an Accidental Reward Cliff

A store proposes a $20 reward when eligible spending reaches $100. At $99.99, the reward is zero. At $100, it becomes $20. The customer adds one cent of eligible value and the proposed benefit increases by $20. That is a reward cliff, and it should be a deliberate feature of the offer rather than an accidental discovery.

A threshold is not inherently a poor design. It simply creates a discontinuity that the merchant needs to understand. The worksheet below examines the mathematical boundary before anyone evaluates real orders or handles later changes.

Plot the reward immediately around the threshold

Define the amount being measured first. For this fictional example, eligible spending is an already-approved merchandise base. The threshold is inclusive: a base of $100 or more earns a fixed $20 reward. Below $100, it earns no reward under this formula. Taxes, shipping, discounts and eligibility decisions have already been resolved outside this calculation.

Use amounts immediately around the boundary, not only convenient examples such as $80 and $120. The one-cent tests expose the precise point at which the amount changes.

The last column is a planning subtraction, not the checkout amount or a claim that reward value is identical to immediate cash. It is included to show the shape of the proposed benefit. The actual original payment remains separate from the later reward.

If the merchant means “more than $100,” the exact $100 row changes. It would earn zero until the next eligible cent. That difference must be explicit in the amount specification. Casual phrases such as “around $100” cannot define a boundary reliably.

Compare the extra spend with the extra benefit

Between $99.99 and $100, eligible spending rises by $0.01 while the nominal reward rises by $20. Between $100 and $110, eligible spending rises by $10 while the fixed reward does not change. Those two comparisons reveal the threshold’s intended incentive pattern.

Do not translate that pattern into a prediction that customers will add products or that revenue will rise. The worksheet describes the reward function, not behavior. Customers may have no useful item to add, may not want the original basket or may prefer another offer. Merchandising and measurement require their own work.

Compare a smooth alternative only to understand the boundary. A 20% reward on the entire approved base would produce approximately $20 at both $99.99 and $100 after rounding, rather than jumping from zero to $20. That alternative changes the whole earning design and its cost, so it should not be treated as a tiny wording adjustment.

Another possible design is a smaller benefit below the main threshold and a larger benefit above it. That reduces or changes the jump but introduces another explicit formula. For example, $5 below $100 and $20 at or above it creates a $15 jump instead of $20, assuming the lower benefit applies throughout the specified range. The merchant should select the shape intentionally, not add exceptions after seeing awkward examples.

The RebateCardX campaign overview is a starting point for discussing an approved threshold concept. The numbers here are original examples and do not establish program limits or available configuration. Present the boundary table when asking whether the selected design is supported.

A difficult case involves rounding the eligible base itself. If an internal amount contains fractions smaller than a cent, the team must know whether the threshold compares a defined cent amount or an unrounded calculation. Do not let reward rounding quietly decide purchase qualification. The commercial specification should identify the authoritative amount before the threshold is applied.

Approve the intended boundary behavior

The approval question is concrete: does the merchant want a fixed $20 increase in nominal benefit when the approved base reaches $100? If yes, the cliff is part of the design. If no, revise the formula before it becomes a promise.

For the fictional store, a completed decision might read: “Approve a fixed $20 reward for an eligible base of $100 or more. No reward is produced below that boundary under this formula. The $99.99, $100 and $100.01 outputs are respectively $0, $20 and $20. The discontinuity is intentional.”

That statement should be accompanied by a precise definition of eligible value from the relevant amount-base decision. A threshold applied to merchandise after discounts behaves differently from one applied to another reference amount, even when both display the number $100. This worksheet does not choose the base; it tests the already-selected one.

Keep the boundary decision separate from product recommendations. Once the threshold is approved, the complementary-product worksheet can assess whether useful additions exist for baskets below it. That later work should not redesign the threshold or encourage unnecessary purchases solely to produce a reward.

Finally, ask two reviewers to calculate the exact-boundary row independently. If one uses greater-than and the other uses greater-than-or-equal, the design is not yet settled. Resolve the difference in the specification, not in a future support case. The completed table gives every later team a stable, inspectable definition of the reward cliff the merchant has chosen.

Discuss the customer-value jump created by your proposed spend threshold. Discuss program fit.

Source references

• RebateCardX campaign overview
• Discuss program fit

Back to contents