How to queue multiple teller lines without customers picking the "wrong" one

Walk into a branch with four teller windows and you'll often see four different lines, or one messy cluster of people trying to watch all four at once. Either way, the customer is the one doing the load-balancing: guessing which line is shortest, which teller is fastest, and whether the person who just walked up to window 3 is about to cut ahead of everyone who's been waiting at window 1.

Why separate lines create an unfair-feeling wait

A line per teller only works if every teller finishes at the same pace, and they never do. One teller is handling a straightforward deposit, another is walking a customer through a dispute that's going to take fifteen minutes. The customer who picked the slow line has no way of knowing that in advance, and no way to switch once they've committed. The actual wait might be perfectly reasonable on average, but it doesn't feel reasonable to the person who picked wrong.

This is a fairness problem, not just a speed problem. People are far more tolerant of a long wait that's clearly first-come-first-served than a shorter wait where they can see someone who arrived after them get served first, just because that person chose a different line.

Example: a Friday afternoon at a branch with four windows

It's 4pm, payday Friday, and all four teller windows are staffed. A customer walks in and splits the difference, standing between windows 2 and 3 where the lines look shortest. Ninety seconds later, someone joins window 1 directly, which was empty because the previous customer there just left; that new arrival gets served before either of the people who'd been waiting at windows 2 and 3 for five minutes. From the window 1 customer's side, nothing went wrong. From everyone else's side, that's the exact moment a reasonable wait starts to feel like a bad one.

One queue, pulled to whichever teller is free

The fix is the same one post offices and the better-run clinics already use: one line, one number, and whichever teller finishes next takes the person who's been waiting longest. A few things make this work well for a teller floor specifically:

1. Let a teller call the next customer the moment they're free, not on a fixed rotation. A teller who just finished a quick transaction should be able to pull the next person immediately, rather than waiting for "their turn" in some round-robin that ignores how fast each window is actually moving.
2. Keep specialty transactions visible but still fair. If certain windows only handle specific services (new accounts, business banking), that's a separate queue by service, not a separate queue per teller doing the same thing. The fairness problem is about identical services split across lines, not genuinely different services that belong apart.
3. Show the actual position, not just "please wait." A customer who can see they're third in line, regardless of which teller eventually calls them, stops needing to track which physical line is moving. The line itself becomes irrelevant to them.

Where Qrua fits

Qrua replaces the per-teller lines with a single QR code queue: customers scan it, get a token, and see their live position and an estimated wait time from their own phone instead of guessing which window to stand near. Staff work from one shared queue and call the next token from whichever window opens up first, so speed differences between tellers help the line move faster overall instead of punishing whoever picked the slower one. If the branch genuinely needs separate lines for different services, like new accounts versus everyday transactions, that's a second queue, not a second guess for the customer.