The Loyalty Card That Lost Money
A Plovdiv chain of three coffee shops launched a loyalty programme: every 10th coffee free, plus a 15% discount on the customer's birthday month. Before launch, average monthly revenue per shop was BGN 28,000 with a 65% gross margin and 4,200 transactions. Six months later, transactions are up 22% to about 5,100 per shop, but revenue is only BGN 29,500 and gross margin has fallen to 54%. The owner, Vesela, believes the programme is a success because footfall rose. Staff report that many customers now buy only during promotions, and that the free coffees are often claimed by customers who would have come anyway. The card also collects phone numbers, but no one has analysed the data. Vesela is considering extending the discount to 20% to 'keep momentum'.
1. Use the numbers to evaluate whether the loyalty programme is actually working.
Read the worked answer
Compare like with like. Revenue rose from BGN 28,000 to BGN 29,500, about 5%, while transactions rose 22%. That means average spend per transaction fell sharply: previously about BGN 6.67, now about BGN 5.78. Gross profit before was 65% of BGN 28,000, about BGN 18,200. Now it is 54% of BGN 29,500, about BGN 15,930. So the programme added transactions but destroyed roughly BGN 2,270 of gross profit per shop per month. Footfall is the wrong metric; the right one is gross profit per member and incremental visits by customers who would not otherwise have come. The free-coffee mechanic rewards existing loyal customers, which is pure margin leakage, and the birthday discount trains customers to wait for promotions. Extending the discount to 20% would likely deepen the margin decline unless it brings genuinely new customers. Vesela should calculate the break-even lift: to restore BGN 18,200 of gross profit at a 54% margin she would need revenue of about BGN 33,700, far above current levels.
2. Recommend a redesign of the programme and the measurement plan.
Read the worked answer
First, change the reward from a blanket free item to something that increases basket value, for example a free add-on or a spend-threshold reward, so the reward does not cannibalise the core purchase. Second, target the birthday offer only to lapsed customers rather than all members, which converts it from a discount for regulars into a win-back tool. Third, segment the phone-number data: identify members by visit frequency and average spend, and compare their behaviour before and after joining, ideally with a control group of non-members in one shop. The key measure is incremental gross profit per member, not transactions. Set a clear test: run the redesigned programme in one shop for eight weeks and compare gross profit per transaction and visit frequency against the other two. Only roll out if incremental gross profit exceeds the cost of rewards. Finally, cap discount depth and avoid raising it to 20%, since the current data show the margin cannot absorb more discounting without a proven new-customer effect.