Spots regular store customers who have quietly stopped coming back
TL;DR Lapsewatch reads a retail chain's billing history and flags regular customers whose visits have slipped past their usual gap. Each store gets a short daily list to call or message. Chains of 5–50 outlets in eyewear, footwear and apparel pay per outlet, and it works if a recovered customer is worth more than the call.
India's retail market was about US$1,094 billion in 2025, and organised chains keep growing their share[1]. Most mid-size chains already collect a phone number at billing, yet hardly anyone looks at who stopped buying.
A customer who bought spectacles every 18 months or school shoes every June just doesn't show up one year, and nobody notices until the year-end numbers dip.
Tools exist but they point elsewhere. One loyalty tool, with 32,000+ restaurants, is built for food service[2]. Loyalty platforms focus on points and campaigns for large brands[3]. So a 12-outlet optical chain in Pune has all the data and nobody to read it.
Lapsewatch connects to the chain's POS or a nightly export. For each customer it learns a normal buying gap from their own history and category, then flags people who are well past it. Store staff get a ranked list every morning with the reason - "usually buys every 7 months, last bill 11 months ago, two pairs of school shoes".
The seed wanted psychographic AI, and we dropped it. Most chains do not hold personality or lifestyle data, and guessing it from shoppers brings consent and notice duties under the DPDP Act[4]. Purchase behaviour alone carries most of the signal, and it's data the chain already uses to serve the customer.
The output is a call list plus a weekly report on how many flagged customers came back and what they spent. Messages go only to customers who agreed to hear from the store. The hard part isn't the model. It's getting a busy store to make five calls a day.
The buyer is the owner or CRM head of a chain with 5–50 outlets. That restaurant loyalty tool charges ₹3,250 per outlet a month on its paid plan[5], which sets a rough ceiling for a single-purpose tool. My working assumption is ₹1,500–3,000 per outlet a month, and that needs testing. With sceptical owners a pilot priced on recovered revenue may close faster.
The DPDP Act passed in 2023 and its rules are now in force, so chains need cleaner consent records anyway[4]. Phone-number billing is close to universal in organised retail. Survival models that once needed a data team now run cheaply on a laptop-sized server, and published work shows a ranked list can catch most leavers - one 2026 banking study caught about 80% of churners in the top 40% of scores[6].
There is no public figure for what Indian retail chains spend on churn analytics. That tool's 32,000+ restaurant customers[2] show that small outlets will pay for repeat-customer tools.
Adjacent - total Indian retail at US$1,093.89 billion in 2025[1]. That's context and not the addressable market.
Restaurant loyalty tools offer RFM segments and win-back campaigns, mainly for restaurants and cafes[5]. Enterprise loyalty platforms run points and customer data for big fashion and beauty brands[3]. Large CRM suites can do lapse scoring if someone sets them up.
The gap is a narrow tool for specialty chains, with category-aware buying cycles and a daily call list a store manager will actually use.
| Difficulty | Moderate (35/100) |
| Starting capital | ₹55–95 L |
| Time to launch | 6–9 months |
The model is the easy part of this build. POS integrations across Indian billing software are messy, and data quality swings from outlet to outlet. The real work is changing behaviour in stores and proving that recovered customers wouldn't have come back anyway.
No sector licence applies. The rules that matter are about customer data and messaging.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't let AI guess personality, income or lifestyle from purchases. It's weak signal and brings consent problems, so stick to what people actually bought.
Want to work out the AI layer for Lapsewatch? Talk to Yash →
Here's what can go south -
I like this for a founder who has sold software to retail chains and doesn't mind sitting in stores. If you want a pure data-science product then skip it. First prove, with a 10% holdout, that called customers come back at least twice as often as uncalled ones.
Plan for roughly ₹55–95 L, mostly on engineers, POS integrations and field sales to chains. Starting with one chain's export keeps integration costs low.
No sector licence. You need company registration, GST, DPDP compliance as the chain's processor and opted-in lists for any messages.
Yes, any market with chain retail and POS data works. Check your local data protection and marketing consent rules before you message anyone.
The free Lapsewatch validation kit has the line-by-line capital breakdown, the effort scoring, the data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for chain owners and store managers, and 5 assumptions to test. Download the kit →
Thinking about building Lapsewatch or something close to it? The make-or-break work is proving recovered revenue to an owner who has seen plenty of dashboards, and that's exactly the kind of thing I help founders with. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.