Fill One Quiet Afternoon Without Losing Lunch Trade

Sofia straightens an untouched table in her Georgetown bistro. Tuesday and Wednesday afternoons are empty from two until five, although staff are rostered and rent is paid. Her previous RM2 lunch discount cost RM12 across six meals that regulars might have bought anyway. It gave nobody a reason to arrive later.

Haziq works nearby and eats out most days, always at lunch. An afternoon meal here has never crossed his mind. Sofia, Haziq and this bistro are fictional; all figures and dates are illustrative. Can Sofia give him a reason to come at three without losing his lunch business?

1. A specific invitation. Haziq receives a message naming Tuesday 15 September, 2pm–5pm, a free iced tea with a RM20 meal, and six available rewards. He can judge whether that afternoon suits him.
2. A believable limit. Haziq claims one of the limited rewards on the dated day. Sofia has genuinely allowed only six, with one claim per customer.
3. A different visit. Haziq arrives at 3pm and buys the meal. Staff check the time and purchase before honouring his reward.
4. A measured next step. In the successful example, Sofia fills six covers in a previously dead hour. Haziq returns to ordinary pricing afterwards; Sofia schedules a fresh window next month only if the test passes.

Sofia chooses Tuesday alone, leaving Wednesday outside this test. The earlier lunch discount has ended: both baseline Tuesdays and the test use the ordinary RM20 lunch price, with RM9 variable meal cost. Her invitation explains that claiming a reward does not extend its hours or reserve a table; the team plans space for all six possible arrivals.

She uses a simple till tally to record arrival time, meal payment and tea served. She also compares lunch covers against similar Tuesdays and asks returning customers whether they shifted an intended lunch visit. Six afternoon diners are encouraging, but she waits until the fortnight’s review before deciding whether the offer added business.

• Net additional covers: quiet-window gain over baseline minus peak visits lost; Sofia’s target is five.
• Peak covers lost: compare lunch trade with matched Tuesdays and customer explanations. If the loss exceeds the quiet-window gain, stop.
• Campaign contribution: collected cash less fulfilment, tea, operating costs and displaced peak contribution. Count every tea served, even on a shifted visit.

Common mistake: calling every afternoon redemption a new customer visit. Haziq changing his lunch time is not automatically growth.

Safeguard: the shift lead checks and records the stated hours and qualifying meal before serving each promotional tea. A claim on the right day is insufficient outside the afternoon window.

Keep when five net additional covers, positive contribution and the safeguard hold. Improve a passing test with a weak signal by changing only invitation timing next time. Stop this version if contribution is non-positive, the safeguard fails or it misses five net additional covers.

Can this run every Tuesday automatically?

No; a repeating weekly happy-hour reward rule is not supported here. Choose a fresh dated window after reviewing the previous test.

What if Haziq claims at lunch or arrives after five?

Claiming does not override the stated hours. Staff honour the offer only during 2pm–5pm on the named date and explain that condition in advance.

Do six claims mean six occupied seats?

No; someone may claim and never arrive. The reward limit controls availability, while Sofia plans seating and measures actual paid covers herself.

What if lunch gets quieter too?

Subtract displaced lunch visits and their lost contribution before judging success. Stop if peak covers fall by more than the quiet window gains, even if every tea was used.