A change order is the estimating team's scoreboard. Every one tells you something you missed on bid day. Some are unavoidable — owner scope changes, differing site conditions. But a meaningful share of COs on a typical commercial project are scope gaps: work that needed to happen, that nobody bid, that got caught only when it showed up on the schedule.
The question worth asking is what a missed scope gap actually costs. The direct cost — the CO amount — is the number everyone talks about. It's also the smallest component.
The four costs of a missed scope gap
1. The direct cost: the CO amount itself.
2. The markup cost: the CO sub knows they have you. They're not bidding competitively — they're quoting a price to fix a problem you discovered after award. Our customer data shows CO markups running 15–40% over what the work would have cost at competitive bid.
3. The delay cost: a missed scope gap almost always comes with a schedule impact. Even a two-week delay on a $5M project with $15K/day in general conditions is $100K in burn before anyone picks up a tool.
4. The relationship cost: the owner's project team learns that their GC's estimate missed something. That's a mark against you on this project and the next RFP.
A worked example
Assume a $5M commercial interior project. During construction, the fire caulking scope is caught as a gap between the drywall sub (who excluded it) and the fire protection sub (who also excluded it).
Direct cost: $18,000 for fire caulking from the subcontractor brought in to handle it.
Markup cost: competitive bid would have been ~$14,000. The $4,000 delta is the tax you pay for catching the gap late.
Delay cost: fire caulking wasn't in the schedule, so MEP rough-in got held up waiting for it. Three days of delay at $8,000/day general conditions = $24,000.
Relationship cost: hard to quantify, but the owner's PM noticed. Next RFP the owner goes out on, you're one of four GCs competing; that reputational ding matters.
Total real cost of the missed scope gap: $46,000+ on an $18,000 CO. Ratio of real cost to direct cost: 2.5x.
This is typical. Our customer data on ~800 analyzed COs suggests the median ratio is 2.1x direct cost when you include markup and delay. The relationship cost is real but doesn't get a number.
Why scope gaps are under-caught
Most GC estimating teams level bids by comparing totals. They ask 'is this bid in line with the others?' rather than 'what did this bid exclude that the others included?' The second question is the one that catches scope gaps, and it requires column-by-column leveling — not total-to-total comparison.
Three structural reasons scope gaps slip through:
• Exclusions are narrative, not structured. Most sub bids have exclusions in a block of text at the bottom of page 2. They don't map cleanly to the leveling sheet unless someone extracts them into rows.
• Scope lives between trades. Fire caulking, blocking, equipment pads, weather protection — the gaps that bite are the ones between two trades who each assume the other is handling it.
• Leveling happens under time pressure. Bid day is not the time to build a careful scope comparison from scratch. If the structure isn't already in place, the shortcut is to compare totals.
How to actually catch them
Three things have to be true to reliably catch scope gaps at leveling time:
1. The leveling sheet has to be structured by CSI subdivision, not by sub. A sub-by-sub sheet tells you who's cheapest. A subdivision-by-subdivision sheet with one column per sub tells you which rows have blanks — and blanks across every sub in the same row is the canonical scope gap signal.
2. Inclusions and exclusions have to be extracted as tagged lists, not left as narrative text. You can't compare 'per plans and specs' to 'excludes fire caulking, blocking, and painting' without structuring the exclusion text into rows.
3. Historical scope gap patterns have to inform the review. Fire caulking, blocking, and equipment pads are scope gaps in 80% of the commercial interior projects we see. If your leveling process doesn't explicitly check for them, you will miss some of them.
The AI-assisted version
PreconIntel's scope normalization extracts inclusions, exclusions, and clarifications as tagged lists automatically, and the leveling sheet highlights rows where every sub has a blank — the canonical scope-gap signal. Historical patterns from your own bid database (not a generic vendor list) inform which gaps to explicitly check for.
What we're not doing is replacing estimator judgment. The software surfaces signals. The estimator decides whether to add a GC allowance, carry the work on the low bid, or push the sub to include it. That judgment call is still the job.
One number to walk away with
The ratio of total real cost to CO direct cost is approximately 2x on most missed scope gaps in commercial construction. Every dollar you avoid in CO value at leveling is about $2 in savings in the real world when you account for markup and delay.
That ratio changes the ROI math on any tool that helps catch scope gaps earlier. If a $10K/year investment prevents one $20K CO per year, it's break-even on the CO line. It's 2x ROI when you include the real cost.