The Cost of a Late Decision: Why Deferred Design Problems Are Never Free Recent
(Part 2 of 3 in a series on why the AE industry’s design schedule is failing projects)
Last week I wrote about why the AE industry’s standard design schedule — which allocates only 15 to 20 percent of total project time to Schematic Design — is set up to fail. The argument was structural: too much is compressed into a phase that is already eroded by firm utilization, resource ramp-up, and the enabling work that renovation projects require before real design can begin.
This week I want to go deeper on the mechanism that makes a compressed SD phase so costly. Because it is not just that problems get created in SD — it is that problems get deferred there, and deferred problems have a very predictable, very expensive escalation pattern.
The cost of change is not constant
There is a principle that every experienced project manager in the design and construction industry understands intuitively, even if they don’t always act on it: the cost of resolving a design issue depends almost entirely on when it is resolved.
In early Schematic Design, a change costs almost nothing. The design exists as concepts, sketches, and preliminary drawings. Reconsidering a structural system, repositioning a core, or reconceiving a building’s relationship to the site requires time at a whiteboard and a follow-up meeting. The investment is measured in hours.
Move that same unresolved question into Design Development and the cost grows significantly. Now the question touches coordinated drawings across multiple disciplines. Revising the structural approach means revising the mechanical coordination. Repositioning the core means updating the egress plan, the electrical distribution, the plumbing chases. What was an hour of conversation in SD is now days of coordinated redraw.
Move it into Construction Documents and the cost grows again. CDs represent a substantial investment of production time across architecture, structural, mechanical, electrical, plumbing, and specialty consultants. A significant design change at 80 percent CDs doesn’t just affect the drawing that shows the change — it ripples through specifications, schedules, details, and coordination drawings across the entire set.
Get the unresolved question to the construction site and the cost escalates to a different order of magnitude entirely. Now the contractor has priced and scheduled against the documents as issued. A change means a formal change order with contractor markup and overhead. It means subcontractor schedule impacts. It may mean materials that have already been fabricated or ordered need to be modified or replaced. It means schedule extensions. In contested cases, it means claims.
Industry experience consistently bears out a cost escalation ratio in the range of 1:10:100 — meaning the same resolution that costs $1 in SD costs approximately $10 in CDs and $100 in construction. The specific ratio varies by project type and the nature of the decision, but the order-of-magnitude relationship holds.
What “deferred” actually looks like
The problems that become construction change orders rarely originate in construction. They originate in an SD phase that didn’t have the time or resources to resolve them — and then travel forward through the project, growing in cost at each phase boundary they cross.
Here are the patterns I see most consistently;
Existing conditions left unverified. On renovation projects, the question of whether existing horizontal and vertical distribution spaces can accommodate new systems is fundamental. It affects structural design, mechanical routing, ceiling heights, and sometimes the entire viability of a systems approach. When SD is compressed, this verification gets deferred — often with a notation that it will be “confirmed in the field” during construction. What that means in practice is that a contractor opens a ceiling and discovers the available plenum is four inches shallower than the drawings assumed. The mechanical contractor submits a change order. The structural engineer reviews rerouting options. The project owner pays for all of it, plus the schedule impact.
Utility capacity assumed, not confirmed. Confirming that existing electrical service, gas supply, chilled water, or domestic water infrastructure has adequate capacity for a project’s new loads is an SD-phase task. When it gets deferred to DD or CDs, the confirmation either never happens or happens late — and a finding of inadequate capacity at 60 percent construction documents is a serious problem. Adding service capacity means utility coordination, potential equipment relocation, additional switchgear or infrastructure, and budget pressure at exactly the moment when the design should be headed toward completion.
Structural assumptions that don’t survive contact with reality. A new rooftop mechanical unit requires structural support. The SD-phase assumption is that the existing roof framing can carry the load with minor reinforcement. When the structural engineer actually analyzes the existing framing in DD — or worse, when the contractor opens the roof during construction — the framing turns out to be significantly lighter than assumed. The mechanical unit has to move, or the structural reinforcement is far more extensive than planned. Either outcome triggers a cascade of design revisions and cost increases.
Stakeholder conflicts that weren’t surfaced. An SD phase that moves too quickly through owner and user group engagement produces nominal approvals — sign-offs from people who weren’t fully engaged with the design direction being presented. When DD drawings are issued and a department head sees for the first time exactly what “open office” means for their group’s workflow, or when a facilities director realizes the mechanical room location will create maintenance access problems, the objection arrives in the wrong phase. Resolving it in DD costs more than resolving it in SD. Resolving it in CDs costs more still. And if it isn’t resolved until construction is underway, the cost is in a different category entirely.
The compounding effect
What makes this particularly damaging is that these issues rarely arrive in isolation. A compressed SD phase that defers multiple unresolved questions sends them all forward together, and they tend to compound.
A mechanical rerouting required by a discovered existing condition affects structural coordination. The structural response affects ceiling heights. The ceiling height change affects egress and lighting. Each issue that arrives from SD triggers a chain of secondary revisions in DD and CDs, each of which consumes fixed-fee hours that were budgeted for refinement work, not problem-solving. The design team absorbs the margin impact. The owner absorbs the schedule impact. The project absorbs both.
This is the mechanism by which a compressed SD phase — even one that appears to save time at the front of a project — produces a worse outcome for everyone by the time the project reaches completion.
The arithmetic of front-loading
The implication of all of this is straightforward: the cost of a rigorous SD phase is not additive. It is substitutive.
An hour spent verifying an existing structural condition in SD is not an extra hour added to the project. It is an hour that eliminates ten hours of DD revision and a hundred hours of construction-phase change order management. An iteration cycle with an owner stakeholder group in SD is not a luxury. It is an investment that eliminates the cost of a scope change at 50 percent CDs.
When viewed through this lens, the industry’s default schedule — which compresses the cheapest phase and expands the expensive ones — is not just a scheduling inefficiency. It is a systematic wealth transfer from the owner’s contingency budget to the contractor’s change order log.
Rigorous SD does not cost more. It costs less, every time — just not in a way that shows up on the SD phase budget line. It shows up at project closeout, in the gap between what the project cost and what it should have cost.
What this means in practice
For owners approaching a new project, the practical takeaway is this: ask your design team to show you not just the phase schedule, but the SD milestone plan. What specific investigations, confirmations, and stakeholder engagements are planned, and by when? What are the known risks that need to be resolved before SD can close?
If the SD milestone plan is thin — a kickoff meeting, a few design iterations, and an owner review — the unresolved questions will be resolved later. And later is always more expensive.
For design professionals, the question is whether the SD phase you are proposing is genuinely sized to the work that needs to happen in it, or whether it reflects a convention inherited from a schedule framework that was never designed around project outcomes.
The cost of a late decision is not hidden. It is entirely predictable. The only question is when and where on the project timeline it will be paid.
Next week: a post addressed directly to building owners, because the downstream consequences of a compressed SD phase aren’t felt most acutely by the design team. They’re felt by the people who hired them.
I’m curious what patterns others have seen — from either side of the table. Where do deferred SD decisions tend to surface most painfully on the projects you’ve been involved with?