The service hall: many desks, one queue

A public service hall has a structural problem that a bank branch does not: a single transaction often requires several desks, in order, and the citizen has no way of knowing which. They queue for the first counter, are told they need a stamp from another, queue again, and are told a document is missing. Two hours in, they have not completed anything.

The queue was never the bottleneck. The sequence was.

Check the documents before the queue, not after

The most effective change most service centres can make is to move validation to the front. A short screening step — at a kiosk, on a phone before travelling, or at a greeter desk — that confirms the applicant has what they need turns a wasted two-hour visit into a two-minute redirect.

It also protects the counters. Staff time spent explaining a missing document is time not spent processing the queue behind it.

One token, several desks

Give each department its own prefix and its own counters, but keep them on one board and one token. A citizen who needs three stops sees all three as one journey, and each desk sees what has already been done.

Practically, that means a token that survives a transfer. When counter 4 finishes, it does not close the token — it hands it to the cashier queue with its history attached, and the citizen is called again rather than starting again.

Priority rules that survive scrutiny

Public services usually have legally or politically mandated priority categories — elderly visitors, people of determination, pregnant women, senior citizens. These need to be handled explicitly:

• Declared at intake, not decided at the counter.
• Bounded, so a priority queue cannot indefinitely starve the standard one. A ratio — one priority call for every three standard — is defensible; “always first” is not.
• Logged, so the reason a token jumped is recoverable months later.

Reporting for compliance, not for reconstruction

Service-level commitments in the public sector are usually published: a maximum wait, a percentage served within a target, an availability window. Meeting them is one problem; evidencing them at the end of the quarter is another, and it is the one that consumes staff time.

Because every token is timestamped at issue, call and close, the report is a query rather than a project. Wait distribution, percentage inside target, volumes by department, exceptions with reasons — exported, not reconstructed.

What to measure

• Percentage served within target, by department, not overall.
• Multi-desk visits. How many citizens need more than one counter, and how long the whole journey takes.
• Failed document rate. High numbers mean your published requirements are unclear.
• Priority ratio actually applied, against the ratio in policy.