Sagar Auto Care: Repeat Visits or New Faces?
Sagar Auto Care is a two-bay two-wheeler workshop in Cuttack with 1,800 names in a paper register that doubles as its customer database. Over the last 12 months the workshop billed 1,050 jobs: 440 were first-time customers and 610 were repeat visits. Only 26% of first-timers return within a year; those who do return average 2.5 visits a year with a contribution of ₹700 per visit, while a first visit contributes only ₹280. The owner spends ₹32,000 a month on Google ads and pamphlets that reliably bring 50 new customers a month. A WhatsApp win-back offer, sent 45 days after a first visit and carrying a ₹150 voucher, would cost about ₹25 per message (sending plus staff time) and is expected to bring back 12% of those customers. Saturday bays are already 80% booked; weekdays run at 55%. The owner has ₹40,000 a month and believes he must choose between advertising and retention.
1. Calculate the cost of acquiring a first-time customer and the expected contribution that customer produces. Is the Google and pamphlet spend worth continuing? Show the arithmetic and name the assumption that most affects your answer.
Read the worked answer
Cost per acquisition (CPA) = ₹32,000 ÷ 50 = ₹640.
Expected contribution from one first-timer = ₹280 (first visit) + 0.26 × 2.5 × ₹700 (later visits) = ₹280 + ₹455 = ₹735.
Net = ₹735 − ₹640 = ₹95 per customer.
Judgement: the spend is defensible but fragile. The first visit alone loses ₹360 (₹280 − ₹640); the ad only pays because of the 26% who return. If the return rate slips from 26% to 20%, expected contribution becomes ₹280 + ₹350 = ₹630, below CPA, and every acquired customer is a loss. So continue only while the return rate is measured (capture the mobile number on the first bill, mark repeat visits), and treat raising 26% as the real target rather than buying more first visits of the same quality.
The assumption that drives the answer is that this month's 50 new customers behave like the historic 26% - same area, same vehicle mix, same price. If the ads are now reaching a cheaper, more distant segment, even ₹95 disappears.
2. Evaluate the WhatsApp win-back offer on its own numbers. Should the owner choose between advertising and retention, or is the choice false?
Read the worked answer
Monthly win-back cost = 50 messages × ₹25 = ₹1,250, plus vouchers: 12% of 50 = 6 redemptions × ₹150 = ₹900. Total ₹2,150.
Contribution from those 6 visits = 6 × ₹700 = ₹4,200. Net = ₹4,200 − ₹2,150 = ₹2,050.
Cannibalisation is limited, because 74% of first-timers would not have returned anyway; most redemptions are visits that would otherwise not exist. Even if one of the six would have come back regardless, the offer still clears its cost.
Judgement: the choice is false. Retention needs ₹2,150 out of ₹40,000, so the owner can run both. Fund the win-back first because it returns roughly ₹4,200 of contribution on ₹2,150 and carries no capacity shock - six extra jobs a month is nothing next to 50 acquired customers. Then keep advertising at a level the bays can absorb. The win-back also raises the 26% return rate, which is precisely the number that makes the ₹32,000 ad spend profitable; the two 'options' reinforce each other.
3. The owner asks you to run the win-back for three months. What will you change about capacity and how will you prove it worked, given that the register is paper-based?
Read the worked answer
Capacity. Six extra visits a month is small, but Saturday is at 80% while weekdays sit at 55%. Make the voucher valid Tuesday–Thursday only and have the caller book a slot while on the phone. This fills idle weekday capacity and avoids displacing full-price Saturday work.
Measurement gap. The register cannot prove anything yet: it records neither mobile numbers for non-returners nor the source of each visit. Before starting, add name, mobile, visit date and 'how did you hear about us' to every job card, and print a unique code on each voucher so redemptions are traceable.
Test design. Split each month's 50 first-timers at random into 25 messaged and 25 not messaged (a holdout). After 90 days compare return rate and contribution per customer, not just voucher redemptions - redemptions only show who took the discount.
Decision rule. If the messaged group's return rate exceeds the holdout by enough to cover ₹2,150 plus the voucher margin, scale it; otherwise stop. Three months gives roughly 75 messaged and 75 holdout customers, enough to see a difference this size.