---
title: 'Hard-bid vs CMAR vs design-build: estimating workflows for GC estimators'
url: 'https://piper-ai.com/resources/hard-bid-vs-cmar-vs-design-build-estimating-workflows'
description: 'Compare how estimating unfolds under hard-bid (DBB), CMAR/GMP, and design-build (solicitation, scope, leveling, carry, and closeout) with checkpoints and checklists for each method.'
contentTypes: [comparison, guide]
publishedAt: '2026-08-06T18:00:00.000Z'
updatedAt: '2026-10-14T08:02:32.643Z'
author: 'Ido Gedanken, CEO'
readingTime: 6
---

# Hard-bid vs CMAR vs design-build: estimating workflows for GC estimators

1. Delivery method changes when design freezes, how you solicit trades, and who carries design vs cost risk
2. Hard-bid is sequential (complete docs, sealed bids, then build) so coverage and addenda discipline decide the number
3. CMAR/GMP is phased: precon advice, trade buyout, then a negotiated GMP with explicit contingency
4. Design-build overlaps design and construction, estimators price incomplete design and iterate with the team
5. Adjust scope review, contingency, and coverage targets by method; document assumptions in every case

**Project delivery method shapes the estimator's workflow.** Hard-bid (design-bid-build) prices a complete set and wins on lowest responsive bid. CMAR/GMP brings the GC in during design, then negotiates a guaranteed maximum price. Design-build puts design and construction under one contract, overlapping phases and putting both design and build risk on the design-builder.

Each method changes how you run solicitation, scope development, bid leveling, the carry / contingency decision, and post-award handoff. Missteps in coverage or assumptions hurt in all three, especially when design is incomplete. This guide compares the workflows phase by phase so GC estimators can adjust checklists, risk, and trade coverage by method.

## Delivery method definitions

### Design-bid-build (hard-bid)

The owner hires an architect for full design, then solicits sealed competitive bids. Separate contracts: designer and lowest-responsive GC (typically lump sum). Flow is linear, **design → bid → build**. Common in public work and wherever competitive bidding is required.

### CMAR / GMP (construction manager at risk)

The CM joins during design as a cost/schedule/constructability advisor, then holds trade contracts under a fixed or **guaranteed maximum price**. Owner keeps separate designer and CM contracts. Design risk largely stays with owner/designer until GMP; after GMP the CM carries construction cost risk up to that ceiling. Trade buyout often starts around 60-90% design. Popular on complex projects that need early contractor input with price assurance.

### Design-build

One design-builder delivers design and construction under a single contract, both design and build risk on that team. Design and construction can overlap (foundations while finishes are still open). Faster delivery and fewer owner-side design COs are the usual pitch; it still needs clear performance criteria and tight design-cost coordination.

## Solicitation and selection

| Aspect | Hard-bid (DBB) | CMAR / GMP | Design-build |
| --- | --- | --- | --- |
| Owner solicitation | Invitation to bid / RFP, price-driven | RFQ/RFP on qualifications (and fee) | RFP on goals, program, budget, criteria |
| When trades are priced | After complete docs | Often ~60-85% design into GMP | Design-builder selects / shops subs; not a public trade bid at award |
| Negotiation | Little on price, responsive to fixed docs | Open-book precon; GMP negotiated | Proposal price (often GMP) negotiated / qualifications-based |
| Estimator focus | Assemble packages, track addenda, hit bid day | Precon advice + phased trade buyout | Conceptual proposal estimate + iterative alignment |

**Hard-bid.** Final price is set against **completed design**. Missing scope becomes a change later. See [solicitation](https://piper-ai.com/resources/construction-bid-solicitation-guide) and [addenda management](https://piper-ai.com/resources/construction-addenda-management-for-gc-estimators).

**CMAR.** Decision point: competitive sub-bids (common on public CMAR) vs negotiated / self-perform (more common private). Either way, trade numbers firm the GMP.

**Design-build.** Upfront cost and schedule commitment often happens on less-than-complete design. Estimators must predict unresolved work and justify it.

## Scope definition and estimating prep

### Hard-bid checklist

1. Verify the final drawing set; flag late design changes
2. Confirm invited and required trades per package
3. Prepare scope outlines and distribute to estimating / PM
4. Set up a [bid coverage matrix](https://piper-ai.com/resources/construction-bid-coverage-guide)
5. Identify long-leads and owner-furnished items in the specs
6. Build the initial cost model with allowances only for minor unknowns

Design should be complete at bid. Incomplete ground reports or missing details become qualifications and risk, not silent optimism. Scope craft: [scope generation](https://piper-ai.com/resources/construction-scope-generation-guide).

### CMAR process

1. At 30-60% design, ROM estimates for major scopes; flag underground and other major risks
2. Track budget vs design with a target-budget report; value-engineer as needed
3. Near 60-85%, issue trade packages; hold allowances / contingency for remaining design
4. Maintain a record of design assumptions and unresolved RFIs / details

Push owner decisions (finishes, systems) while design is still open. Use placeholder allowances consciously. They feed the [carry number](https://piper-ai.com/resources/carry-number-methodology-construction-bids).

### Design-build checklist

- Lock performance criteria with the owner (budget, start date, capacity, energy)
- Compile historical $/SF or system benchmarks for similar DB work
- Structure major packages (foundation, shell, MEP, finishes) that may run in parallel
- Set design freeze / allowance cutoffs (e.g. finishes locked at 30% as deduct/credit, refined at 60%)
- Document scope decisions as the living baseline, iterative design without a register becomes dispute fuel

## Bid and trade package management

### Hard-bid

Trades and quantities are locked at bid time. Use a coverage matrix so every critical package has enough qualified bids:

| Trade package | Sub A | Sub B | Sub C | Notes |
| --- | --- | --- | --- | --- |
| Concrete | $350k covers | $370k covers | - | Sub C excluded basement slab |
| Structural steel | $200k covers | - | $220k covers | Sub B no-bid |
| Electrical | $150k | $140k | $145k | Scope aligned across bidders |
| HVAC | $180k | $175k | - | Only two bidders; thin coverage |

Keep an [addenda log](https://piper-ai.com/resources/construction-addenda-management-for-gc-estimators). On bid day: level scopes item-by-item, chase high-impact [exclusions](https://piper-ai.com/resources/review-subcontractor-exclusions-and-qualifications), hunt [scope gaps](https://piper-ai.com/resources/spot-scope-gaps-before-you-carry), and decide whether thin packages need another invite or a plug ([coverage](https://piper-ai.com/resources/construction-bid-coverage-guide), [leveling](https://piper-ai.com/resources/construction-bid-leveling-guide)).

### CMAR / GMP

- Trade bidding at the design threshold: fewer bidders or direct negotiation possible on private work; public owners often still require competition
- Open-book buyout: normalize allowances and unit prices across evolving scope
- Keep a **scope reconciliation log**: what each trade assumed until final design
- GMP = accepted trade costs + CM markups + contingencies + owner allowances, with contingency justified
- Uncovered items (every mech bid excludes a valve) still need an allowance in the GMP, same discipline as coverage, less formal packaging

### Design-build

- Internal packages; early mini-RFPs to preferred or shortlisted subs for gut-check pricing
- Full competitive rounds may wait until later design stages
- Estimator owns proposal scope completeness against owner promises
- Track changes from the proposal baseline in a living estimate, site or slab shifts become documented cost movements, not surprises

## Final estimate, carry, and risk allocation

| Factor | Hard-bid | CMAR / GMP | Design-build |
| --- | --- | --- | --- |
| Design completeness | Full set | Partial at GMP | Early concept → iterative |
| Who carries design risk | Owner / designer (Spearin) | Owner until GMP; then mostly construction risk on CM | Design-builder |
| Who carries cost risk | GC (except claimable gaps) | CM up to GMP; owner above (with exceptions) | Design-builder |
| Typical contingency posture | Often 2-5% if design truly complete | Often 5-10% when GMP is early | Often 5-15% or padded into line items |

**Hard-bid.** Carry covers residual unknowns (soils, weather, permits) and known gaps. Run [final bid review](https://piper-ai.com/resources/final-bid-review-qa-qc-general-contractors) before submission.

**CMAR.** Contingency is explicit in the GMP narrative. Use a factor-based rubric (design completeness, geotech, scope clarity, market volatility, owner change propensity, complexity). Method detail: [carry number methodology](https://piper-ai.com/resources/carry-number-methodology-construction-bids).

**Design-build.** Contingency is often implicit in line items; some proposals show an owner contingency separately. Because the design-builder cannot bill the owner for its own design mistakes, pad against unresolved refinements with eyes open.

## Handling incomplete design and scope gaps

- **Hard-bid**: force clarifications via RFI/addenda; every ambiguity gets a conservative assumption. Use historical CO trends (e.g. local K-12 \~3% post-bid growth) to size risk when docs are imperfect.
- **CMAR**: design-gap checklist by spec section; regular design reviews; allowance drawings with unit prices or lumps for unresolved scope before GMP lock.
- **Design-build**: design-freeze thresholds and a scope-status register; drift past allowances triggers review before it becomes buried cost.

## KPIs by method

| KPI | What to watch |
| --- | --- |
| Cost growth / accuracy | Research often shows design-build with lower average cost growth than CMAR or DBB, partly because DB absorbs its own design errors. Track bid vs model and final vs GMP/bid. |
| Schedule growth | DB typically fastest; DBB often sees larger schedule overruns. Bid logs and addenda turnaround matter more when the path is compressed. |
| Bidder response | Hard-bid target often 2-3+ bids per trade. CMAR/DB may run leaner, fewer bidders usually means higher risk premium in the carry. |

Treat published averages as directional; build your own internal medians by delivery method.

## Common failure modes

| Method | Failure mode | Mitigation |
| --- | --- | --- |
| Hard-bid | Incomplete scope / sole-source specialty | Cross-trade gap review; coverage matrix; chase thin packages early |
| Hard-bid | Late owner addenda flood | Buffer time; addenda log; liquidated-damages awareness |
| Hard-bid | Suspiciously low bid | Full exclusions pass before you celebrate the number |
| CMAR | A/E-CM disconnect | Standing coordination; reconcile drawings to buyout assumptions |
| CMAR | GMP surprise when trades land high | Lock major equipment / civil early; transparent contingency narrative |
| CMAR | Owner adds scope pre-GMP expecting inclusion | Owner-changes allowance; adjust GMP explicitly |
| Design-build | Vague performance brief | Written performance specs before price commitment |
| Design-build | Friendly but expensive preferred subs | Benchmark against market |
| Design-build | Permit / AHJ late demands | Early plan-reviewer contact; freeze thresholds |

## Trade-specific notes

- **Structural, earthwork, utilities**: quantities fixed in DBB; still moving in CMAR/DB until site design settles. State quantity assumptions and allowances clearly.
- **MEP**: DB/CMAR can value-engineer systems early (chiller type, distribution). DBB focuses on small omissions against a complete spec.
- **Finishes and specialty**: stage long-lead millwork; lock colors/fixtures later with a small contingency in CMAR/DB; in DBB, price architect allowances explicitly.

## Spreadsheets vs dedicated tools

| Situation | Usually enough | Prefer a dedicated system |
| --- | --- | --- |
| Fully designed DBB, smaller jobs, disciplined files | Spreadsheet leveling and summaries | Optional |
| Partial design (CMAR/DB), fast-track, many estimators | Painful, version drift and missed links | Single source for bids, RFIs, assumptions, and cost model |

The test is traceability: can you show which document or sub proposal backs each number when design or addenda move? That is the same "source-backed" bar as leveling and final review, method does not change the need; incomplete design raises the stakes.

## Choosing a method (owner lens, estimator impact)

1. **Lowest price on complete docs** → hard-bid, estimator focuses on responsive coverage and a clean bid day
2. **Budget certainty plus contractor input during design** → CMAR/GMP, estimator lives in precon updates and a justified GMP
3. **Speed and single point of responsibility** → design-build, estimator iterates design and cost together and prices unresolved work up front

## Phase timing (illustrative)

| Phase | Hard-bid | CMAR / GMP | Design-build |
| --- | --- | --- | --- |
| Design / precon | Design complete before bid | Design + precon services overlap for months | Proposal development, then overlapping design-construction |
| Trade pricing | Hard bid date after docs | Trade bids into GMP (~60-85% design) | Internal / staged buyout as design advances |
| Price commitment | Final sealed bid | GMP negotiation | Proposal / GMP with iterative refinement |
| Construction start | After award | After GMP (often) | Can start early packages while design continues |

## Where Piper fits

Every row in the comparison above is a workflow the estimator runs anyway: define the scope, get bids against it, absorb the addenda, and review the number before it leaves. What the delivery method changes is the order and the tolerance, not the work.

Piper is the AI operating system for preconstruction. It turns the bid set into trade scopes you can hand out, compares the proposals that come back against those scopes rather than against each other, and checks the estimate against the project behind it before it goes out, keeping every finding attached to the clause or sheet it came from. The same project understanding carries between those steps, so an assumption made while the documents were still incomplete is on the screen when the addendum that tests it arrives.

Piper does not run your solicitation, set your contingency, or decide which method an owner should use. It keeps scope, bids and the estimate reading from the same project so the differences between methods stay deliberate.

### Does hard-bid always mean lower contingency?

Only when design is truly complete and coverage is strong. Incomplete docs or thin trades need the same honest carry you would use on CMAR, silence is not a method advantage.

### When should CMAR start trade bidding?

Commonly around 60-85% design, enough detail to get real numbers, enough openness left that allowances for remaining design are still explicit in the GMP.

### Why do design-build estimates feel "softer" early?

Because you are pricing a proposal against performance criteria and incomplete design. The rigor is in freeze points, living baselines, and benchmarking, not in pretending the set is 100%.

### How many bids per trade by method?

Hard-bid: target 3-5 usable bids (invite more). CMAR/DB: often fewer, especially with preferred teams, document sole-source rationale and raise contingency when competition is thin.

### Where do exclusions reviews still matter on CMAR and DB?

Everywhere you take a trade price. Open-book does not remove exclusions; it makes normalizing them part of the GMP story. Same matrix discipline as [exclusions review](/resources/review-subcontractor-exclusions-and-qualifications).