This page teaches how to test whether a loyalty reward changes behaviour, for an agent, a brand-side contact, or a buyer deciding whether to approve a proposed reward. This lesson mainly teaches the decision flow: it compares a customer's own buying behaviour with a reward against the same customer's behaviour without it, so a business can tell a reward that genuinely drives extra buying apart from a reward that simply reduces profit on purchases that would have happened regardless. A business that rewards every customer equally, without checking which customers' behaviour the reward actually changed, can spend real money on a reward that changed nothing.
You need: an understanding of repeat-customer value against acquisition cost, covered in Lesson 4.3, and access to purchase data for a group of customers who received a reward and a comparable group who did not.
By the end of this lesson you can:
Not covered here: which customer data fields a loyalty programme actually needs (Lesson 1), segmenting customers by real purchase behaviour (Lesson 2), calculating repeat-customer value against acquisition cost (Lesson 3), and building a complete customer data and loyalty worksheet (Lesson 5).
| Term | Meaning |
|---|---|
| Behaviour-changing reward | A reward that causes a customer to buy more, or buy sooner, than that customer would have without it. |
| Behaviour-neutral reward | A reward paid to a customer who would have made the same purchase anyway, adding cost without changing what the customer buys. |
| Comparison group | A group of customers who did not receive a reward, used to see what buying behaviour looks like without it. |
| Incremental purchase | A purchase a customer makes because of a reward, over and above what that customer would otherwise have bought. |
| Reward cost | The total value given up by a business through a reward, whether as a discount, a free item, or points redeemed. |
| Net reward value | The extra profit a reward produces from incremental purchases, minus the reward's own cost. |
A reward paid to a customer who would have bought the same thing anyway is a cost with no benefit attached; only the purchases a reward genuinely causes, incremental purchases, justify its own cost. Comparing a rewarded group's own buying behaviour against a comparison group that received no reward is how a business tells a behaviour-changing reward apart from a behaviour-neutral reward, rather than assuming every reward works the same way.
| What can differ | The shapes you meet | Why it matters |
|---|---|---|
| Export hybrid via an agent vs. domestic MTS retailer | An export hybrid business selling through an agent such as Meridian Sourcing can test a reward across separate European retail partners as natural comparison groups; a domestic MTS retailer such as Cedar & Co. can split its own shop customers by branch or by sign-up date instead | The comparison group has to be genuinely similar to the rewarded group; comparing very different customer types risks blaming or crediting the reward for a difference that already existed |
| A reward tested on a small trial group first | A business offers a reward to a small, clearly defined group and compares it against a similar untested group before rolling it out further | A small trial limits the cost of finding out a reward is behaviour-neutral, before committing to a full rollout |
| A reward rolled out to every customer at once | A business gives every customer the same reward immediately, with no comparison group left | Without any comparison group, a business cannot tell whether the reward changed behaviour or simply paid out to purchases that would have happened anyway |
| A reward targeted at customers who already buy frequently | A reward aimed at a retailer's own most frequent existing customers | Frequent existing customers are often the least likely to change behaviour because of a reward, since they were already going to buy; a reward aimed at occasional or lapsed customers can show a clearer behaviour change |
| Product families | The same behaviour-change test applies whether the reward is offered on apparel, socks and hosiery, knit-to-garment, terry and towels, or home textiles and bedding | A business new to this test should build its own comparison group from its own real customers, not assume another product family's result applies |
Meridian Sourcing, an agent in Egypt working on a 5% commission for two European buyers, proposes testing a 10% discount on a customer's next order, offered to a group of Northline Apparel's own smaller European retail partners that reorder the Order A T-shirt style. Half of these partners, chosen at random, receive the discount offer; the other half, a comparable group in size and past order pattern, receive no offer and continue as usual, serving as the comparison group.
Over the following quarter, the rewarded group places reorders averaging 1,850 pieces per partner; the comparison group, with no reward, places reorders averaging 1,600 pieces per partner. Northline Apparel's own European sales manager wants to know whether the discount actually caused this difference, and whether the extra profit it produced was worth the discount's own cost.
The worked numberReading the comparison group's own 1,600-piece figure is what lets Northline Apparel separate the 1,600 pieces the discount did not need to buy from the 250 incremental pieces it did. Once the discount's own cost on only those incremental pieces is subtracted, the reward still produces a positive net value of USD 120 per rewarded partner, showing this specific reward genuinely changed behaviour rather than only paying out to reorders that would have happened anyway.
A common mistake is looking at a rewarded group's own total buying and crediting the whole amount to the reward, without checking what a comparable unrewarded group bought over the same period. If Northline Apparel credited the reward with the rewarded group's full 1,850-piece average, rather than only the 250-piece incremental gap the comparison group reveals, it would badly overstate the reward's own value and could keep funding a reward that mostly pays for purchases it never actually caused.
A reward is only worth its cost if it makes a customer buy more than they would have anyway, and the only way to know that is to compare a rewarded group against a similar group that received nothing. On Meridian Sourcing's own test, the rewarded group ordered 250 more pieces per partner than the comparison group; after subtracting the discount's own cost on just those extra pieces, the reward still earned back USD 120 more per partner than it cost, showing it genuinely changed behaviour rather than only paying for orders that would have happened regardless.
Northline Apparel's own European sales manager has received a proposal from the sales team to extend the 10% discount to all European retail partners next quarter, citing the rewarded group's strong 1,850-piece average reorder. The facts in front of the sales manager:
The note attached to the sales team's own proposal:
Should the sales manager approve the proposal based on the rewarded group's own total average alone? Decide before you open the decision.
No. The sales manager should approve the rollout based on the calculated net reward value of USD 120 per partner, not on the rewarded group's own total average alone.
Two facts decide this. First, this lesson's own rule states a reward is only worth its cost for the incremental purchases it causes, and the worked number shows only 250 of the 1,850 pieces are incremental; the rest would likely have been ordered anyway, based on the comparison group. Second, this lesson's own trap warns exactly against crediting a reward with a rewarded group's full total; the proposal's own justification does precisely that, citing the 1,850-piece average without mentioning the comparison group at all.
The rollout may still be worth approving, since net reward value is positive at USD 120 per partner, but that approval should rest on the calculated incremental figure, not on the rewarded group's own total average, which overstates what the reward actually achieved.
Correctly measured, the reward's own net value is USD 120 per partner. Crediting it instead with the rewarded group's full order overstates the order value driven by the reward by USD 800 per partner, a gap that would compound each time this mistake is repeated across future reward decisions.
Use the T23-C04 Customer data and loyalty worksheet. Enter the rewarded group's own average order and the comparison group's own average order, along with the reward's own cost rate, and the sheet calculates the incremental purchase, its own profit, the reward's own cost on that incremental amount, and net reward value.
[Download the Customer data and loyalty worksheet](../../templates/T23-C04-customer-data-and-loyalty-worksheet.xlsx)
1. Why is a reward only worth its cost for the incremental purchases it causes, not for a rewarded customer's total buying?
2. Meridian Sourcing's rewarded group averages 1,850 pieces; the comparison group averages 1,600 pieces. What is the incremental purchase the reward caused?
3. A retailer tests a reward on 100 customers and finds the rewarded group and a similar comparison group both bought the same average amount. What does this tell the retailer?
| EN | AR | FR | TR | Meaning |
|---|---|---|---|---|
| Behaviour-changing reward | مكافأة مُغيّرة للسلوك | Récompense modifiant le comportement | Davranış değiştiren ödül | A reward that causes a customer to buy more, or buy sooner, than that customer would have without it. |
| Behaviour-neutral reward | مكافأة محايدة للسلوك | Récompense neutre sur le comportement | Davranışa nötr ödül | A reward paid to a customer who would have made the same purchase anyway, adding cost without changing what the customer buys. |
| Comparison group | مجموعة المقارنة | Groupe de comparaison | Karşılaştırma grubu | A group of customers who did not receive a reward, used to see what buying behaviour looks like without it. |
| Incremental purchase | الشراء الإضافي | Achat incrémental | Artışsal satın alma | A purchase a customer makes because of a reward, over and above what that customer would otherwise have bought. |
| Reward cost | تكلفة المكافأة | Coût de la récompense | Ödül maliyeti | The total value given up by a business through a reward, whether as a discount, a free item, or points redeemed. |
| Net reward value | صافي قيمة المكافأة | Valeur nette de la récompense | Net ödül değeri | The extra profit a reward produces from incremental purchases, minus the reward's own cost. |
You now know how to test whether a loyalty reward actually changes a customer's own behaviour, rather than simply paying out to a customer who would have bought anyway. The final lesson of this course brings every tool from Lessons 1 to 4 together into one complete customer data and loyalty worksheet. Back to the course map