Simulator and golden parts

This article shows you how to test a change to your pricing before any buyer is quoted from it. The panel beside the configuration editor prices a real part from your draft, live, as you type. It is the same engine that answers buyers and the same configuration you are editing, just not published yet.

Use the simulator

1. Drop a part on the panel beside the configuration editor.
2. Set a quantity and choose options as a buyer would.
3. Change a rate and watch the breakdown move. Change a quantity break and see whether the price curve looks like the one you meant to draw.

This is the fastest way to answer “what would this do?”, and it is much better than the alternative: publishing and finding out from an invoice.

Golden parts

A golden part is a job whose right price you already know, because you quoted it by hand and won it, or because you have made it a hundred times. Pin it with the price you believe in, and Quotometry checks every draft against it.

The value is that you see all of them at once. A change that fixes one job and quietly breaks three is the most common way a pricing model goes wrong, and it is nearly invisible unless something is watching for it.

Which parts to pin

Pick parts that disagree with each other:

• the smallest job you are willing to take, the one your minimums exist for;
• a typical job, your bread and butter;
• a big one, where machine time dominates;
• a batch, where the quantity break does the work;
• an awkward one: heavy supports, a thin wall, a material you rarely use.

Five parts like that catch far more than twenty similar ones. How many you can pin depends on your plan.

The backtest

On plans that include it, the backtest re-prices up to your last 500 quote lines with the draft, and shows how the prices would move and which quotes move most. Use it for a wider check before a big change; use the simulator and golden parts for the change you are making right now.

The order to work in

1. Calibrate first. There is no point tuning rates against a model that is 20 % out on machine time: you would only build a compensating error into your margin.
2. Then pin golden parts and look at the gaps that remain. Those gaps are commercial decisions (margin, minimums, quantity breaks), not estimation errors.
3. Then publish.

Next steps

• Calibration how-to: fit the model to your machines first.
• Drafts, versions and publishing: what happens when you publish.
• Formula language: rules that a form cannot express.