Little's Law Example for a Service Business (Cleaning)

By JJ Andrade, Business Performance Engineer. Six Sigma Black Belt and author of “Queuing Theory for Business Optimization”. Published October 2026.

Most Little’s Law examples online use an airport security line or a bakery. This one uses a small cleaning company, because that is where many owners feel the problem: customers wait too long for a first visit, and nobody can say why. All company names and numbers below are fictional and chosen to make the math easy to follow.

What is Little’s Law in plain English?

The average number of jobs in your system equals the rate at which jobs finish times the average time each job spends waiting and being done.

The formula is short:

L = λ × W

L = average number of jobs in the system (booked, waiting or in progress). This is your work in progress, or WIP.
λ (lambda) = average throughput, the number of jobs finished per unit of time.
W = average time a job spends in the system, from the moment the customer asks until the job is done.

John D. C. Little published the proof in 1961. It works for any stable system, as long as you measure all three numbers over the same period and with the same boundaries. It does not need any assumption about how customers arrive.

How does Little’s Law apply to a cleaning company?

In a cleaning company, L is the open job list, λ is cleanings finished per workday, and W is how many days a customer waits from request to clean.

Think of the whole company as one box. A request enters the box when a customer asks for a cleaning. It leaves the box when the cleaning is finished. Everything in between counts: the request sitting in the inbox, the quote, the waiting for a free slot on the calendar, and the cleaning itself.

That gives you three numbers you can measure with a notebook or your scheduling app:

Letter	What it means here	How to measure it
L	Open jobs: requested but not yet finished	Count open jobs every morning for 2 to 4 weeks and take the average
λ	Jobs finished per workday	Count finished cleanings per workday over the same weeks and take the average
W	Days from request to finished clean	Calculate it with L ÷ λ, or measure it from request dates and finish dates

What is a worked Little’s Law example for a cleaning company?

A fictional company with 36 open jobs that finishes 9 jobs per workday makes each customer wait 4 workdays on average: 36 ÷ 9 = 4.

Meet “Example Cleaning Co.”, a fictional company. All numbers are made up for this example.

Step 1. Measure throughput (λ). The company has 3 crews. Each crew finishes 3 cleanings per workday. So the company finishes 3 × 3 = 9 jobs per workday.

Step 2. Measure open jobs (L). The owner counts open jobs every morning for four weeks. Some mornings the list has 33 jobs, some mornings 39. The average is 36 open jobs.

Step 3. Solve for wait time (W). Little’s Law says L = λ × W. Rearranged, W = L ÷ λ.

W = 36 ÷ 9 = 4 workdays

So a new customer waits, on average, 4 workdays from the moment they ask until their home is clean. The owner did not need a stopwatch or software to find this. Two counts and one division were enough.

Step 4. Check the units. λ is in jobs per workday, so W comes out in workdays. If you measure λ in jobs per week, W comes out in weeks. Mixing units is the most common mistake.

How do you use Little’s Law to cut customer wait times?

Wait time only drops if open jobs fall or throughput rises. To cut W in half at the same throughput, the open job list must be cut in half too.

Say the owner of Example Cleaning Co. wants customers to wait 2 workdays instead of 4. Demand stays at 9 new requests per workday.

L = λ × W = 9 × 2 = 18 open jobs

So the target is clear: keep about 18 open jobs on average instead of 36. Little’s Law tells you the target. It does not tell you how to get there. Common ways to get there are:

Clear the backlog once. Add temporary capacity (an extra shift or a weekend crew) until the list drops from 36 to about 18.
Keep capacity a little above demand. If the crews finish exactly 9 jobs per day and 9 new requests arrive per day, any bad day (a sick cleaner, a cancelled van) rebuilds the backlog. A small buffer keeps it down.
Limit work in progress. Set a cap on how many jobs can be open at once, and give new customers an honest start date instead of an open-ended promise. I explain how in WIP limits: why less work in progress means more output.
Cut time that is not cleaning. Slow quotes and slow scheduling count as time in the system too. They add to W even when the crews are fast.

What happens to wait times when demand grows?

If new requests arrive faster than crews can finish jobs, the open job list grows every day and wait times keep climbing until you add capacity.

Little’s Law describes a stable system: on average, jobs leave as fast as they arrive. Now say a good month of marketing pushes demand to 12 requests per workday, but the 3 crews still finish only 9.

Each workday, 12 jobs come in and 9 go out. The open list grows by 3 jobs per workday.
After 10 workdays, the list has grown by 3 × 10 = 30 jobs, from 36 to 66.
If the company could finish 12 jobs per day, a 66-job list would still mean 66 ÷ 12 = 5.5 workdays of wait. At 9 jobs per day, it keeps growing.

The lesson: more marketing without more capacity does not bring more cleanings. It brings a longer line. Before you add demand, check that λ can follow.

What does the summary table look like?

Same fictional company, three situations. L = λ × W holds in every row, so if you know any two numbers you can find the third.

Situation (fictional)	Throughput λ (jobs per workday)	Wait W (workdays)	Open jobs L
Today	9	4	36
Goal: shorter wait	9	2	18
Demand at 12, capacity raised to 12, same 4-day promise	12	4	48
Demand at 12, capacity stays at 9	9 out, 12 in	keeps rising	grows by 3 per workday (not stable)

You can test your own numbers with the free Little’s Law and M/M/c calculator.

What are the limits of Little’s Law?

It works with long-run averages in a stable system. It will not predict one customer’s wait or tell you how many crews you need at peak hours.

Keep these points in mind:

Averages, not promises. If the average wait is 4 days, some customers wait 1 day and some wait 8. To size crews for peaks and variability, you need a queuing model such as M/M/c. The calculator includes one.
Same boundaries for all three numbers. If L counts quotes waiting for approval, W must include that time too.
Stable period. Measure over a few normal weeks, not during a holiday rush or right after a big promotion.
It does not find the cause. It tells you how big the wait is. Finding why the list is long (a bottleneck in scheduling, quoting or one crew) is the next step. See the invisible queue problem and how to reduce bottlenecks.

How can I calculate this for my own service business?

Count open jobs each morning, count finished jobs each workday, average both over a few weeks, then divide open jobs by jobs finished per day.

A simple 4-week routine:

Every morning, write down how many jobs are open (requested but not finished).
Every evening, write down how many jobs were finished that day.
After 4 weeks, average each column.
Divide average open jobs by average jobs finished per day. That is your average wait in workdays.
Compare it with the wait you promise customers. If the math says 4 days and you promise 2, the gap is the problem to solve.

This works for plumbers, HVAC techs, pet sitters, lawn care and any business that books jobs. The words change; the math does not.

Where can I learn more about queuing in service businesses?

Start with the related articles below, then the book “Queuing Theory for Business Optimization”, which covers these models in more depth.

If you want help measuring L, λ and W in your own operation, you can book a free 30-minute operations diagnostic.