Every heavy-civil GC has done it: the structural concrete package comes back with the low number from a sub you would not put on a critical-path activity if you had a choice. You know the award should go to the second number. The problem is never the decision — it is the record. When a passed-over bidder calls your VP, or an owner’s rep asks why the line item ran 9% over the low quote, “we just had a better feeling about them” is not an answer. A defensible award is built the same day you make it, not reconstructed three weeks later. Good award memo software exists for exactly this gap: it captures the override, the scored trade-off, and the page-cited evidence while the reasoning is still fresh.
Low is not the same as responsible
The lowest bid is only the right award when it is also responsive (it covers your scope) and responsible (the sub can actually perform). On private heavy-civil work you are not bound by public low-bid statutes the way a DOT is. You are free to award on value — but freedom cuts both ways. With no statutory shield, the burden of justification lands on you. The number that looks lowest on the tab sheet is frequently the highest once you account for what it leaves out.
The usual reasons a low bid is not responsible:
- It is low because it is incomplete. The sub omitted dewatering, or excluded the temporary shoring your spec requires, or carried a smaller quantity than your SOW. The delta you are “saving” is a change order waiting to happen.
- The unit prices are unbalanced. A nominal $1.00 on a major pay item, or a mobilization line carrying 14% of the total, tells you the bid is structured to get paid early or to game a quantity overrun — not to reflect cost.
- The bidder is a peer outlier. When one number is less than half the peer median on the same item, that is not a sharp pencil. It is a scope misread or a mistake you will eat mid-job.
- Performance and capacity don’t support it. A sub with no crew availability in your window, an open quality dispute, or a bond they cannot post is not responsible at any price.
Lowest-responsible bid vs. best value
These are two different award standards, and confusing them is how defenses fall apart. Under lowest responsible bid, you award the cheapest bid that clears a pass/fail responsibility gate — price is the only ranking variable once a bidder is “in.” Under best value, price is one weighted factor among several, and a higher number can win on the strength of schedule, scope completeness, or risk.
Both are defensible. What is not defensible is running best-value logic in your head while telling everyone you used low-bid. Pick the standard before you open the bids, state it, and score against it. A clean way to frame it:
| Dimension | Bidder A (low) | Bidder B (awarded) |
|---|---|---|
| Base price | $2,140,000 | $2,318,000 |
| Scope coverage | 3 items missing | Full SOW |
| Schedule fit | 2 wks past milestone | On milestone |
| Risk flags | 2 (unbalanced, mobilization) | 0 |
| Adjusted position | +$165k once gaps priced | Lowest true cost |
Once you price the three missing items into Bidder A, the “low” bid is no longer low. That is the whole argument, and it is one you can put in writing. For a fuller treatment of the two standards and when each applies, see the companion article linked below.
What actually makes an award defensible
A defensible best-value award has four elements in the record. Miss one and the decision looks like a preference instead of an analysis.
1. A named override category and a written reason
If you are passing on the low number, you are overriding the default. Name why in a fixed category — scope gap, schedule risk, performance history, unbalanced pricing, compliance — and write one specific sentence under it. “Awarded B over low bidder A: A excluded sheet-pile shoring (Spec 02260), pricing the gap adds ~$165k and erases the spread.” That sentence is your defense in compressed form.
2. Scored trade-offs, not adjectives
Score every bidder on the same dimensions — price, scope, schedule, compliance, performance, risk — so the comparison is apples to apples and the winner is the output of a method, not a vote. The score is what converts “we liked them” into “they ranked highest on a published rubric.”
3. The scope gaps the low bidder missed
Run a coverage check that lists exactly which SOW items each bid omitted and which extras it added. “Three items missing” is a claim. “Items 0204.02 dewatering, 02260 shoring, 0555.01 temporary striping, not present in Bidder A’s submission, page 4” is evidence.
4. Deterministic risk flags
The strongest flags are rule-based, not judgment calls, because no one can argue with arithmetic. Four that carry weight in a challenge: unit prices at or below $1.00; any item more than 2× or less than half the peer median; a total bid more than 20% off the field; and mobilization above 10% of the total. When you can say “the low bid tripped two of four deterministic risk rules,” the conversation is over.
Build the record at decision time, not after
The single most common failure is reconstructing the rationale weeks later, when the tab sheet is buried, the markups are gone, and you are quoting numbers from memory. By then you cannot cite the page, you cannot prove the gap, and the audit trail has a hole in it.
Build it the day you decide:
- State the decision mode in writing before opening bids — lowest responsible, best value, lowest risk, schedule priority, or a custom weighting — so no one can claim the criteria moved to fit the winner.
- Capture every bid’s line items normalized to your SOW, so the comparison is on common items and the gaps surface automatically.
- Log the override category, the reason sentence, and the page citation at the moment of the call.
- Freeze a snapshot: scores, flags, gaps, and the recommended winner, with whatever you overrode and why. That snapshot is the document you hand up when the decision is questioned.
If you want a fast way to pressure-test a bid before you commit, the free bid-risk scorecard walks the same four risk rules over a single bid and shows you where it is exposed. It is a useful framing device in the award meeting itself — it turns “this one feels risky” into a list of specific, citable flags.
What to do when the award is challenged
Challenges usually come from one of three directions: a passed-over sub, an owner or program manager asking why a line ran over the low number, or your own internal review. The response is the same regardless of source — you produce the record, you do not relitigate the feeling.
- Lead with the standard. “This package was awarded best value, stated before bids opened.” That alone defuses the “but they weren’t the cheapest” opener.
- Show the adjusted price. Walk through the scope gaps you priced into the low bid. When the low number stops being low, the challenge loses its premise.
- Cite the flags by rule and page. Deterministic rules and page citations are not opinions. They are reproducible.
- Produce the override audit trail. The category, the reason, the timestamp. A forced override that was logged at decision time reads as diligence. The same override reconstructed later reads as a cover story.
The difference between those last two is the entire game, and it is decided by when you wrote it down, not by how well you write under pressure.
Use the scorecard as a framing tool, not a verdict
A risk scorecard does not pick your sub — you do. What it does is force the conversation onto specific, named exposures so the decision is argued on evidence. Score the field on the six dimensions, run the four risk rules, list the scope gaps, and the defensible award usually selects itself. When it does not — when you still want to override the highest-scoring bid for a reason the model can’t see — that is fine, as long as you log the override and its reason. The audit trail does not require you to follow the score. It requires you to be honest about when you didn’t, and why.
That is the whole discipline: decide on a stated standard, score the trade-offs, capture the gaps and flags with citations, and log every override the moment you make it. Do that and the award memo writes itself — and the next call from a passed-over bidder is a two-minute conversation instead of a two-week file hunt. A tool built to generate the reasoned award memo from that same record turns the discipline into a default rather than a chore.