The certain cost of a permanent discount is the price reduction given to customers who would have bought anyway. Its benefit is the genuinely additional business it attracts. That does not make every evergreen offer a mistake: a clearly defined member group, partner funding or useful off-peak fence can make one pay. The question is whether the additional contribution exceeds the ongoing markdown.

Who uses the offer you leave running?

An existing customer may have the eligible bank card and claim the offer on every visit. A first-time visitor may return—but still use the same discount. Neither case automatically earns the merchant a full-price repeat. Before renewing an offer, separate first-time visitors, existing full-price customers and repeat offer users in the POS or booking record. Ask the programme who controls dates, exclusions, redemption limits and funding; do not assume a bank pays for the markdown.

“Always on” also needs care as a data label. In DiscGo's 28 September 2026 index, 53% of 31,293 distinct deals had no recorded end date. That means the catalogue lacked a date; it does not verify that those offers run indefinitely. About 16% stated a day or time limit, based on approximate keyword rules.

How many new customers make 20% off pay?

Take an illustrative, merchant-funded Dubai café offer. A check is AED 63 including VAT, or AED 60 before VAT. Ingredients, packaging and payment costs are assumed to total AED 18, leaving AED 42 contribution at full price. A 20% discount removes AED 12 before VAT and leaves AED 30 contribution from a genuinely new customer. It costs AED 12 when used by someone who would have paid full price.

The break-even share of redeemers who must be additional is AED 12 ÷ AED 42 = 29%, rounded. Suppose the café records 7,200 redemptions in a year:

Those results assume capacity is available and every additional check has the same cost. If ingredient and other variable costs are higher, the break-even share rises. If a partner actually funds part of the discount, it falls—but verify the contract and tax treatment first.

Does a standing deal change the price customers expect?

It can, but the long-term evidence is not one-sided. Experiments and panel studies from non-UAE consumer-goods markets found that discount frequency can affect reference prices (Kalwani and Yim, 1992; Mela and colleagues, 1997). Another study found lasting promotion effects largely absent in the products it examined (Pauwels and colleagues, 2002). None proves how a Dubai café's customers will react. Lead with the measurable markdown cost; treat any brand or price-expectation effect as a hypothesis to test.

An evergreen benefit makes more sense when it is limited to a group you intend to serve differently, applies mainly to otherwise empty capacity, is genuinely co-funded, or rewards later behaviour rather than cutting every bill. A broad offer at all hours deserves stronger evidence than a Tuesday-morning one.

Decision rule: Estimate the additional share through a branch or time-period holdout. Keep the offer only if that share exceeds discount rate ÷ full-price contribution margin—about 29% in this example—and the result survives all channel and funding costs. If you cannot show that, narrow the eligible group or hours, or give the offer an end date. Check the applicable promotion rules before advertising; do not infer permit requirements for a specific cardholder deal from this article.