Change-order recovery model

Every buyer request joined to the variance order it triggered and the invoice that settled it, with the scope that was built and never billed named by home and by trade. Bill what you built, at the policy you already wrote.

$1.1M

Reference-scale value

Data sets in the model

6 weeks

Build and deploy

Derived

Evidence grade

The decision this model makes

A buyer asks for something after contract. A superintendent says yes, because saying no costs a relationship. Somebody writes it up later, or does not. The trade invoices for it either way.

Two hundred and seventy-nine of four hundred and fifty buyers take at least one change order. Eight and a half per cent of that scope is built and never billed.

Margin on what does get billed runs 11.4% against a policy of 26%, which is the second gap and the one nobody has measured, because change orders are priced by whoever is standing there.

The model joins every buyer request to the variance order it triggered and the invoice that settled it, and names the home and the trade on every unbilled line.

Cause 01

Unbilled scope

Built, never reached an invoice. The only line in the segment that turns into cash inside a quarter.

Cause 02

Priced below policy

Billed at 11.4% against a 26% policy. Part competitive, part nobody knowing the policy.

Cause 03

Variance orders received

Buyer-initiated variance orders handed here by the purchasing model rather than claimed there.

Cause 04

Signed after the work

Priced at whatever the trade had already charged, because the negotiating position was gone.

Want more detail?

The rest of this model — the data sets it runs on, the cause split, the working behind the reference figure and a sample output — goes out by email. Two fields, nothing else.