Part 1 mapped the quote-to-cash process at Example Windows Co., a made-up window maker. An order takes 66 days to become cash, and about $1.06 million sits in open orders. This post lets you test changes to that process before you spend money on them.
The model below is a working simulation. Orders arrive, wait, pass through five machines and get paid. The results change when you change the settings.
The simulator
Every tile or window frame is one customer order worth $20,000. Its colour shows where the money is. Start with the buttons in the Try row, then move the sliders.
Open the simulator on its own page for a larger view.
How to read the floor
The floor follows the order of the process. Work starts at the back right, runs through the line in the middle and ends at the bank at the front left.
Table 1. What the colours mean
| Colour | Where the money is | Where you see it |
|---|---|---|
| Grey | Quoted, nothing spent | Quote desk and the customer's decision |
| Orange | Cash spent, not billed | Materials, the line, delivery and delivered orders waiting for an invoice |
| Purple | Billed, not paid | The block of unpaid invoices |
| Green | Paid | Tiles moving to the bank |
The purple block is the largest group on the floor. At the starting settings it holds about 64 orders, or $1.28 million of unpaid invoices.
Each machine has a light: green when it is running, yellow when it is waiting, blue during a changeover and red when it has stopped for repair.
Three experiments
1. Invoice on delivery day
Click Invoice on delivery day. The pad of delivered orders waiting for an invoice empties. Quote to cash falls from 66 to 61 days and about $96,000 of cash is freed. No machine changed.
2. Halve the batch size
Click Halve the batch size. Orders stop waiting for a batch of four, so they reach the line sooner. Deliveries on time rise from 80% to 97%. Quote to cash moves by one day and the cash tied up barely changes.
3. Machine trouble
Click Machine trouble. Uptime drops from 90% to 75% and window frames queue in front of the first machine. Only 15% of orders now arrive on time. Quote to cash rises by just 3 days.
Table 2. Results of the experiments
| Setting | Quote to cash | Cash tied up | Delivered on time |
|---|---|---|---|
| Baseline | 66 days | $1,061,000 | 80% |
| Invoice on delivery day | 61 days | $966,000 | 80% |
| Halve the batch size | 65 days | $1,049,000 | 97% |
| Machine trouble | 69 days | $1,098,000 | 15% |
| Fix the office | 50 days | $386,000 | 83% |
The pattern is consistent. The factory decides whether the customer gets the order on time. The office decides how long the money takes. Fix the office combines a one-day quote, same-day invoicing, follow-up of late payers and a 30% deposit, and it frees about $675,000 without touching a machine.
What the model assumes
Table 3. Inputs of the model
| Item | Value in the model |
|---|---|
| Orders | About 1.6 a day, each 20 windows worth $20,000 |
| Costs | $8,000 of materials bought when the order is confirmed, $4,000 of labour added on the line |
| Customer's decision | About 3 days after the quote is sent |
| Order entry and materials | About 7 days, varying from 5 to 9 with the supplier |
| The line | Five machines, 0.4 of a day per order each, with room for two orders between machines |
| Stoppages | Random, about 2 hours to repair |
| Batches | Orders wait for a full batch. The first machine has one changeover per batch. |
| Delivery | 2 days, promised within 15 days of the order |
| Payment | 30-day terms, paid about 10 days late on average |
The model has limits. It shows won orders only, so it cannot tell you how many more quotes a faster answer would win. It ignores weekends and bad debts. Each change restarts the model and runs one year in the background, so the results are settled averages. They come from a random simulation and move a little from run to run.
Use your own numbers
Set the office sliders to your own quote turnaround, invoice lag, payment terms, late days and deposit. The order value and costs stay those of the example. Read the direction and the relative size of each change, not the exact dollars.
Takeaway. Test a change in the model before paying for it. In this example the office settings move the cash, and the factory settings move the delivery promise.
Example Windows Co. and its figures are an illustrative example. The days and dollars come from a simulation model with round-number inputs, and they do not describe any real company.
