Representative case study · MyBiller

How a Mirpur cable TV operator collected from 1,800 subscribers without losing track

The business — call it "Mirpur Cable Network" — serves around 1,800 households across four blocks of Mirpur 12 and 13. Monthly bill is Tk. 400 for standard, Tk. 600 for HD. The owner runs it with three field collectors and a mostly-retired uncle who handles the books.

The bookkeeping system, until early 2026, was: three hardback notebooks (one per collector area), one master ledger at the office, and a WhatsApp group titled "Collection Updates" where collectors sent voice notes after each round.

How the old system broke (every month, in the same way)

Bill cycle is 1st – 7th of each month. Collectors go door-to-door, mark paid in their notebook, hand cash to the office on the 8th, the uncle transcribes everything into the master ledger.

The recurring failures, in order of pain:

1. The "I paid him" dispute. 8 to 12 times a month, a subscriber called to say they had paid but their cable was disconnected. The collector might or might not remember. The notebook entry might be missing, on the wrong line, or written in the wrong month's section. Resolution required physically meeting the collector and walking through three pages.
2. The cash-on-hand reconciliation gap. Collector hands in Tk. 18,400. Notebook says he collected from 47 subscribers. 47 × 400 = 18,800. Where's the Tk. 400? Investigation usually took half a day. Conclusion was almost always "one subscriber paid Tk. 600 not 400 because they're on HD and the collector forgot".
3. The forgotten new connection. When a new household signed up mid-month, their entry would be added to the back of a notebook. About once every two months, a new subscriber was missed for their first billing cycle entirely.
4. The lost notebook. Twice in 2025, a collector lost his notebook. Once it was found. Once it wasn't. The owner had to manually reconstruct around 280 collections from cash deposits, WhatsApp voice notes and customer calls. It took 11 days.

What else the owner tried before MyBiller

• An Excel-on-phone "system": tried for two months. Failed because the three collectors couldn't reliably sync the file. Multiple versions diverged within a week.
• A bespoke web POS-style system quoted by a local freelance: Tk. 65,000 setup + Tk. 2,000/month hosting. The owner walked away after the first call because the demo required constant internet — and his collectors do half their rounds in 3rd-floor stairwells without signal.
• Hire a fourth collector to specialise in dispute resolution. Considered, then rejected: a fourth salary is Tk. 14,000/month; the actual problem was the notebooks, not the collector headcount.

MyBiller was chosen because each collector ran it on their own phone, sync was manual (after the round), and the customer database was the same for everyone — so no version conflicts.

The 3-week rollout

Week 1 — load the subscriber master

The uncle dictated subscriber records from the master ledger; the owner's son typed them into the app: name, phone, address (block + flat number style), monthly bill amount, plan (standard / HD), connection date. Took 22 hours over the week, broken into evening sessions.

Subscribers were grouped into four collection routes matching the existing collector territories. The fourth route exists because of how the boundary between blocks works; the owner had been meaning to clean this up.

Week 2 — one collector, parallel to paper

The youngest collector (and the only one with an Android smartphone he was already used to) was given the app for week 2. He kept his notebook too. After each visit, he marked paid in both.

End-of-week reconciliation: the app showed 312 collections; the notebook showed 314; the cash reconciled to within Tk. 200. The two missing entries turned out to be subscribers the collector had marked paid in the notebook by habit before remembering they'd asked for an extra day.

Week 3 — all three collectors, paper for emergencies

The other two collectors got the app on their phones. The older collector (mid-50s) needed two evenings of one-on-one training. The third collector, who had been with the operator for nine years, was the most resistant — not because he found the app hard but because he didn't trust having his collections visible to the owner in real time.

This was a real, valid concern and the owner addressed it directly: the goal of the app is not to monitor collectors, it's to settle disputes with subscribers. He committed in writing (in the WhatsApp group) not to use the app for surprise audits. After this, adoption was complete.

What changed in the first 90 days

• "I paid him" disputes: 8–12 per month → 1–2 per month. The remaining cases are now resolved in under three minutes by pulling up the subscriber's payment history.
• Cash reconciliation: half a day per collector per round → 15 minutes per collector. The app's collection total matches the cash by default.
• Missed new connections: ~1 per 2 months → zero. New subscribers get added to the app at signup; their first bill is automatic.
• Collection rate (within billing window 1st–7th): 78% → 87%. Not because of any new feature — because the collectors now know exactly who hasn't paid yet, every day.
• Total monthly take: up about Tk. 18,000 / month. Not from new subscribers — from previously-uncollectable arrears that now get visible flags.
• PDF receipts on WhatsApp: zero → about 220 per month. Subscribers like having a receipt. Some had asked for years.

"I thought the app would help me know who paid. The bigger thing is now I know who didn't."
— composite reflection from the owner

What we'd do differently

1. Address the "monitoring" concern publicly, on day one

The third collector's resistance was avoidable. If the owner had set the rules — "we'll never look at individual collection timing, only at totals" — in week one of week one, the wariness wouldn't have built up.

2. Print a per-route subscriber map for each collector

The app's collection list is text. The collectors think in geography. A printed one-pager per route, showing each subscriber's location on a hand-drawn map, would have helped the older collector orient faster in week 2.

3. Set up reminder messages for unpaid subscribers earlier

MyBiller can send a "your bill is due" message via WhatsApp Web. The owner only enabled this in month three. Turning it on at end-of-week-1 of any bill cycle would probably have pushed collection rate up another 3–4 percentage points.

The transferable lessons

1. Field collection apps are political, not technical. Collectors interpret real-time visibility as surveillance unless you explicitly say otherwise. Get this conversation out of the way before week 1.
2. Reconciliation problems are usually data-entry problems. The cash-on-hand gaps weren't theft, they were typos. Putting the data entry on the phone (with hard-coded plan amounts) removes 90% of those gaps.
3. Don't try to "perfect" the customer master before going live. The Mirpur operator's master had about 80 records that were wrong (wrong phone, old address). They got corrected naturally as the collectors visited each subscriber for the first round.
4. WhatsApp receipts are a marketing feature, not just a billing feature. Subscribers like getting them. They also share them, which is informal advertising the operator wasn't paying for.

For ISPs and cable operators specifically, the Bangladesh Telecommunication Regulatory Commission's subscriber record-keeping guidelines set the formal minimums. MyBiller's data export is BTRC-compatible (CSV with the required fields), which matters at licence-renewal time.