Representative case study · Visitor Desk

How a 12-storey Dhaka apartment complex retired its paper visitor register

It's a 12-storey building, 8 flats per floor, 96 units. We'll call it "Lake View Heights" — Dhanmondi area, mostly owner-occupied, a few rented. There's one security gate, two guards on rotation (morning and evening), and a paper register on a clipboard.

The register has been there since the building opened in 2014. It records: visitor name, flat number they're visiting, time in, time out, signature. About 180 entries per week on average. By 2025 it had filled 27 hardback notebooks.

Why the committee finally moved

Three incidents in eight months, building on a slow-burn frustration:

1. March 2026 — courier dispute. A flat owner claimed a package was never delivered. The guard insisted it was. The register entry for that day was unreadable (rain had hit the clipboard during a power cut). No camera footage covered the gate desk. The owner ended up replacing the package out of his own pocket.
2. April 2026 — overstay incident. A domestic worker entered for one flat at 10 AM, was found in a different (unoccupied) flat at 4 PM. The register had her arrival but no exit. Nothing else.
3. May 2026 — complaint about Friday-evening crowd. The committee chair wanted to know how many visitors entered between 6 PM and 9 PM on Fridays. The register had this data buried across nine pages. Nobody had time to count.

None of these were individually catastrophic. Together, they made the committee call a meeting.

What the committee considered

• Hire a "society management" SaaS (Tk. 1,200/month for the building). Bundled visitor + billing + complaints. Killed by two factors: required every owner to install an app and create an account (60% of owners are over 55, half refused), and dropped to read-only when the building wifi was down.
• Buy a biometric / RFID gate. Quoted Tk. 1.8 lakh + Tk. 30,000/year maintenance. Killed by cost and by the fact that visitors (not residents) can't be enrolled in advance.
• Just keep the register and buy a waterproof one. Briefly considered. Killed by the second incident — paper has no overstay alert, period.
• Use Visitor Desk on one tablet at the gate. Chosen. Tk. 14,000 for the tablet, free for the app's standalone tier, no per-owner installs required.

The 4-week rollout

Week 1 — install + train, both registers running

One Lenovo tablet (10 inch) mounted on the gate desk on a swivel arm. Visitor Desk installed, building name and committee chair phone set up. No accounts needed for the guards.

For the entire first week, every visitor was entered in both the paper register and the tablet. This doubled the guard's per-visitor time from 30 seconds to about 90 seconds. They didn't love it. But it built the data needed for the next phase.

The guards were given a one-page printout: how to add a visitor, how to mark exit, how to look up a recent visitor. The morning guard caught on in two days; the evening guard took five (he's 62 and had never used a tablet).

Week 2 — paper for evening shift only

The morning guard was confident enough to drop the paper register. The evening guard kept doing both. By end of week 2, the tablet had 320 visitor records and the morning shift was 100% digital.

Week 3 — overstay alerts on, owner notifications off

The overstay alert was turned on with a 4-hour threshold: if a visitor signed in 4+ hours ago and hasn't signed out, the tablet flashes the entry red. The morning guard noticed and asked about it within a day; the policy of "ring the flat after 4 hours of no exit" was set.

Owner-side notifications (SMS to the flat when their visitor arrives) were deliberately left off at this stage. The committee wanted the gate workflow stable first.

Week 4 — paper retired entirely

The evening guard had been doing both for three weeks and was ready. The paper register went into the committee chair's cupboard "for backup". It hasn't been opened since.

Owner notifications were enabled at end of week 4 — but only for flat owners who had asked for them (37 of 96 flats). The committee was clear that opt-in was the only fair policy.

What the data started showing

Within 60 days the committee had answers to questions they'd never been able to answer before:

• Friday 6 PM – 9 PM visitor count: 42 average. The peak that the original complaint was about. Now visible, debatable, fixable.
• Most-visited flat: 4B (the chair's flat, ironically). 14 visitors / week. Cue laughter at the committee meeting.
• Repeat overstay-pattern visitors: 3 specific names. Two were a cleaning agency the building uses; the third was unknown and led to a brief, awkward conversation with a specific flat owner.
• Average visitor dwell time: 47 minutes. Useful for capacity planning (when to bring the second guard on).

"We didn't need digital. We needed to be able to count. The digital was just how we got the counting done."
— composite reflection from the committee chair

What we'd do differently

1. Run the paper register for only 4–5 days, not a full week

The guards hated week 1. Less of it would have been kinder, and the data confidence was already there by day 4.

2. Get the chairperson to actually log in to the web admin on day 1

The committee chair only opened the web admin in week 5. Doing it on day 1 — even just to verify a couple of records — would have closed the trust loop faster and would have caught one early misconfiguration (a flat number typo).

3. Discuss owner notifications BEFORE rollout, not after

SMS notifications to flats are a contentious topic. Some owners want every visitor logged; some find it intrusive. Resolving this in the committee meeting before rollout, rather than retrofitting opt-in/opt-out lists later, would have saved a meeting.

The transferable lessons

1. The gate workflow has to get faster, not slower. Any digital tool that lengthens the per-visitor time at the gate will be sabotaged by the guards within a month, deliberately or not.
2. Don't require resident installs in residential buildings. Apartment-wide app rollouts fail roughly 100% of the time. Older residents won't install; younger residents resent the data collection. Run digital at the gate; communicate to residents the old way (paper notice + WhatsApp group).
3. Owner notifications are a policy decision, not a feature flag. Decide it in writing, with the committee, before turning anything on. SMS to wrong number = complaint within an hour.
4. Keep the paper register for 60 days after cutover, then archive it. Not for use. For peace of mind during the first dispute.

For multi-tenant residential buildings in Bangladesh, the Real Estate Development & Management Act 2010 sets the legal framework for committee-managed common areas — and the standard of "reasonable care" in disputes increasingly assumes documented visitor logs. The relevant Ministry of Housing & Public Works guidance pages are worth a read for committees that want to understand the standard.