Open change orders in the forecast: weighting what is not yet signed
At any point in a GMP job there are change orders that exist but are not executed. Proposals under review, items priced but disputed, requests the owner has not answered. Together they might be worth $1.2 million on a fictional $45 million project, and none of that is in the GMP yet. The anticipated final cost has to say something about them, and whatever it says is a judgment.
The judgment is unavoidable. What can be avoided is making it differently each month, or making it in a way nobody can see. A weighting by status, set once and applied every month, turns the judgment into a rule that the report can show and the board can question.
The statuses and their weights
A request that has been raised but not priced is the most uncertain: it might be nothing, it might be a real change. A proposal that has been priced but not reviewed is more likely than not to become a change at something near its price. A proposal the owner has reviewed and accepted in principle, pending execution, is nearly certain. A proposal the owner has rejected is not in the forecast at all, unless the contractor has kept it open as a claim, in which case it needs its own line.
A common weighting is 25 percent for raised, 50 percent for priced, 90 percent for accepted in principle, and 100 percent for executed. The numbers are not magic. What matters is that they are written down, applied to every item, and changed only by a decision the report records.
The cause changes the weight
Status is not the only thing that predicts whether a proposal becomes money. Cause does too. A priced proposal for an owner scope change the owner requested in writing is going to happen at close to its price. A priced proposal the contractor has classified as unforeseen and the owner reads as a contractor risk is going to be argued, and may become nothing. The weighting can carry a second dimension, cause, or the owner can simply override the status weight on disputed items with a stated reason. Either way the override is on the record.
Pricing that has not settled
A proposal priced at $180,000 that the owner expects to settle at $120,000 is weighted on which figure? The honest answer is the owner's expectation, with the contractor's figure shown beside it. The forecast carries $120,000 at the status weight; the report shows both numbers so the board sees the range. When the change executes at $135,000, the register records that too, and over a project the pattern of asked against settled becomes a fact about this contractor the owner can use.
Keeping the weighting honest
The danger with a weighting is drift. A bad month arrives and the temptation is to weight a few large items down to make the projection look better. A good month arrives and the temptation runs the other way, to bank the improvement. The defense is that the weights are set by status, the status is a fact about where the item sits in the process, and the month's report shows every item with its status, its value and its weighted value. A member of the board who wants to check can add the column.
The other defense is the stored snapshot. If each month's forecast is kept as it was, with its weighted items, the owner can look back and see whether the weighting predicted well. On the fictional project, if the weighted total of pending items has run at $600,000 and executed changes have averaged $650,000 a quarter later, the weighting is about right. If executed changes have run at twice the weighted figure, the weights are too low and the board has been told a projection that was systematically optimistic.
What the report shows
The pending items appear in the report as a table: number, description, cause, status, value asked, value expected, weight, weighted value. The total of the last column is the term in the anticipated final cost. Below the table, one line says what the projection would be if every pending item executed at its asked value, and one says what it would be if every disputed item were rejected. Those two lines are the range, and the weighted figure sits inside it. The board sees the number and the two edges, which is more honest than any single figure could be.
What to do this month
- Write down the status weights once and put them in the first report that uses them.
- Enter every pending item with its status, its asked value and the owner's expected settlement value.
- Record any override of the status weight with a reason, on the item.
- Compare the weighted total from three months ago with what actually executed, and adjust the weights only by a recorded decision.
Questions on this
What weights should be used?
Any that are written down and applied consistently. A common set is 25 percent raised, 50 percent priced, 90 percent accepted in principle. The consistency matters more than the numbers.
Are rejected proposals in the forecast?
Not unless the contractor has kept them open as a claim. Then they belong on their own line, weighted by the owner's view of the claim and shown separately from ordinary pending changes.
How does the owner know the weighting is right?
By looking back. If each month's forecast is stored as it was, the weighted total from a quarter ago can be compared with what executed since. A pattern of under prediction means the weights are too low.
In the product
Anticipated final cost, Change order register, Monthly owner report. Free tool: Change Order Exposure.
Keep reading
Earlier: The first monthly report: what to show when the project is one month old. Later: Packages bought late in the job: what the last awards do to the savings pool. All articles on change orders.
See the baseline on a contract like yours.
Thirty minutes on a call. We read a fictional GMP amendment into the baseline and show what the basis flags reveal.