How to Tell If Your Office Is Meeting Its Wait-Time Target

Measure the share of served customers helped within your target, then check what that number hides: how long the late customers waited, who walked out, and which queues let down the most people.

Most offices have a wait-time target, even if nobody wrote it down. It might sit in a service agreement, a performance plan, or a line on the website that says most customers are seen within 15 minutes. The hard part is knowing whether you kept it.

What does it mean to meet a wait-time target?

Your wait-time target is the longest wait your office promises customers, for example 15 minutes. Queue management reports usually call the matching measure service level: the share of served customers whose wait ended within the target. An agreement often states both parts together, such as 85% of customers seen within 15 minutes.

A customer's wait starts when they take a ticket and ends when their service starts. Each served customer's wait is checked against the target and counted as in time or late.

Why can't the average wait tell me?

The average blends every customer into one number, so it hides the people who waited too long.

Take ten customers who waited 3, 6, 8, 9, 10, 12, 14, 18, 19 and 33 minutes. The average is 13.2 minutes, under a 15-minute target. But three of the ten waited longer than 15 minutes, and one waited more than twice as long. The average says you kept your promise. Those three customers would say you didn't. For more on this, see why average wait time often misleads operations teams.

How is it calculated?

Count the served customers whose wait was within the target, and divide by all served customers. With the same ten customers and a 15-minute target, seven were helped in time. That's 70%. Customers who walked out are not in this number, which matters more than it seems (see below).

The target changes the score:

The customers waited exactly as long in all three rows. Only the target moved. That's why the target should come from your service agreement, not from whatever makes the number look good.

What is a good percentage?

Start with what your agreement promises. If it says 85% within 15 minutes, 85% is your line.

If there's no written number, these are the bands we start from:

These are starting points, not a rule. A quick pickup counter and a complex case desk can't be held to the same line. Set the bands before you look at the results, and make sure everyone reads them the same way.

What does the percentage hide?

The share tells you how often you kept the promise. It leaves out three things.

How long the late customers waited

Once a wait passes the target, the score counts it as late, whether the customer waited 16 minutes or 90. Two offices can both score 70% while one made its late customers wait 20 minutes and the other made them wait an hour. Check P90 wait next to it: the wait that 90 out of 100 customers stayed under.

The two measures are linked. "90% within 15 minutes" is the same test as "P90 at or under 15 minutes." See P90 wait time: the metric government operations is missing. A score with no wait check beside it can also be gamed: calling customers who are about to go late ahead of customers who are already late raises the score while the longest waits get worse.

Who walked out

The score counts only customers who were served. If many people give up and leave, the ones who stayed may still have been helped quickly, and the score can look fine on a bad day. Check the serve rate next to it: the share of customers who arrived and were actually helped. A good score with a falling serve rate is not a good day.

How many people each queue let down

A queue at 40% looks worst until you see it served 5 customers that week. A queue at 74% that served 4,000 customers let down 1,040 people. Rank queues by how many customers missed the target, not by the percentage alone.

Which queues and days need help first?

Start with the queue that let down the most customers. Look at when its late waits happened, and how busy staff were at those times. In many offices staff serve several queues, so look at their time across those hours, not at one queue alone.

• Staff busy almost all the time, and waits still too long: you need more capacity, or transactions are taking longer than planned. Look at staffing levels and transaction time.
• Staff often idle while customers wait too long: the people are there, but not at the right counters or hours. Look at schedules and which services each person can take.

Then look day by day, next to how many customers came:

• A dip on a busy day: fewer customers were helped in time on a day when many came. Check which queues fell behind that day.
• Below your line on most days: the target may be tighter than the team can meet with current staffing. Talk about the target and staffing before chasing single bad days.

How Meridian shows this

Meridian, the queue analytics product from VelocityNex, reads the data your existing queue system already collects. Your queue system can already print a service level percentage. Meridian puts the context next to it: serve rate and P90 wait beside the score, each queue with its customer count and P90, and each day with how busy it was. A target slider shows what target your current staffing can actually hold, which helps when you need to show leadership what it would take to meet the agreement. It changes only your view, never the agreement. Data refreshes nightly.