Cost exposure
Late design changes
A change instructed in week three of a build costs a fraction of the same change instructed in week thirty. The difference is not the work. It is everything the work now disturbs.
The situation
Two small changes, both arriving after the sequence was set
The design was signed off. Packages went out, prices came back, the contract was executed. Then the client saw the show flat mock-up and wanted the joinery detail revised. Meanwhile the structural engineer issued a revision to accommodate a plant relocation nobody had priced.
Neither change looked significant on its own. However, both landed after the contractor had sequenced the works, ordered materials and committed subcontract packages. What arrives on the client's desk six months later is not the cost of the joinery. It is the cost of the joinery, plus the abortive work, plus the resequencing, plus the preliminaries running while the revised drawings are produced.
By then the client has no leverage. The contractor is on site, the programme is under pressure, and the commercial conversation has become a negotiation about a claim rather than an assessment of a change.
What we found
Change was being recorded, but not valued
Instructed changes were being recorded, but they were not being valued. Consequently the client's cost report showed a contingency drawing down without anyone able to explain which specific decision had consumed it. Three patterns were present, and all three are common.
Changes were issued verbally on site and confirmed in writing weeks later, which removed any contemporaneous record of the commercial position at the point of instruction.
Time impact was being treated as a separate conversation from cost impact, so the same event was assessed twice and agreed once.
Several items described as variations were in fact design development, which the contract already priced. Nobody had tested which was which.
What changed
Valued at the point of instruction
Each instruction was valued at the point it was issued, against the contract mechanism rather than a negotiated rate. The client received a running exposure position rather than a monthly cost report that arrived after decisions had already been taken.
Value on the day, not at the end
Every instruction was priced against the contract mechanism when it was issued, while the site facts were still visible and the client still had the option of declining it.
Separate variation from design development
Where a change was genuinely a variation, it was agreed quickly and closed. Where it sat inside the contractor's existing obligation, that was stated early and evidenced. The distinction decides whether the client is paying at all.
Assess time and money together
The programme consequence was established alongside the cost consequence, so one event produced one assessment rather than two negotiations running months apart.
Close each item before the next month
Nothing was carried forward. The final account contained no accumulated backlog of unvalued instructions to be argued about at the end.
Why this happens
The gap between two different processes
Cost plans are built during design. Change control is applied during construction. On most projects those are two different processes run by two different people, and the handover between them is where exposure enters.
Effective cost control and variations closes that gap by valuing change as a live commercial event rather than a retrospective claim. Furthermore, where quantities are checked against the model as design develops, many late changes are identified before they become instructions at all, which is the function of quantities and take-off.
Questions
Asked about change after tender
Why is a change after tender so much more expensive?
Because competition has ended. Before tender a change is priced by several parties who want the job; afterwards it is priced by one who already has it. The direct cost is only part of the difference. A late change also disturbs procurement already placed, work already sequenced and sometimes work already built, and each of those consequences is valued under the contract rather than tested against the market.
How should a change be valued if it disrupts other work?
Separately from the direct work, and with the disruption evidenced as it happens. Most standard forms distinguish the value of the varied work from the consequential effect on the remainder, and the second is far harder to establish retrospectively. Records taken at the time, showing planned against actual resource, are what make that claim assessable. Where the records are assembled months later from memory, the assessment becomes an argument about credibility rather than about cost.
Can late changes be prevented, or only managed?
Most can be prevented, because most are not changes of mind. They are decisions that were deferred, coordination that was not completed, or information that arrived after the documents were issued. Tracking which of those are still open at tender, and what each is worth, converts a category of surprise into a managed list. What remains after that is genuine change, and it is a much smaller number than most projects experience.
Should we instruct a change before the price is agreed?
Where the works cannot wait, sometimes there is no alternative, but the position should be recorded deliberately rather than by default. Instructing first and pricing later transfers the whole of the commercial leverage, because the work is done and the argument is about what it was worth. Where an instruction has to be issued ahead of agreement, setting out the basis of valuation in the instruction itself preserves at least part of the position.
At what point do accumulated changes need renegotiating as a whole?
When the pattern of change has altered the job the contractor priced. A long run of individual variations, each reasonable on its own, can amount to a different scheme, a different sequence and a different resource profile. At that point valuing each in isolation understates the effect and invites a global claim later. Recognising the threshold early, and dealing with it openly, produces a better outcome than either party arriving at it through a final account.
Related scenarios
Where this connects
Change arriving faster than anyone can value it?
Send the instruction log and the last cost report. We will tell you where the process has broken and what it costs to leave it.
Insights