How to Build a Design Review Rubric That Shortens Meetings

The Two-Hour Design Review That Reviews Nothing

Most design review meetings in design-build engineering don't actually review anything. They perform the review live, in front of everyone, while the schedule clock burns. A scored rubric, completed before the meeting, is what changes that.

You've sat in this meeting. Six people open the design package for the first time at 9 AM. The senior structural engineer interrupts the architect's walkthrough. The MEP lead pulls up a clash she found that morning. Two hours later, three issues are documented and four are forgotten. No one scored anything. It feels like drinking out of the fire hose, which is what the meeting actually is.

"The meeting itself is not a time to actually review the design— it is a time to come with your redlines, questions, concerns, and comments." — Bliley Technologies¹

The macro problem is well-documented. Atlassian's 2024 State of Teams research, covered by Fortune², found meetings are ineffective at disseminating information, encouraging collaboration, and accomplishing tasks 72% of the time. Three symptoms tell you yours is one of them:

• Everyone reviewing the package in real time at the conference table
• Senior people interrupting before scoring is captured
• No written record of which criteria are weak versus which are strong

This article gives you the instrument to fix that: a scored rubric you can use on Monday, and a 30-minute reconciliation meeting that replaces the two-hour group review. The cost of a bad review is higher in design-build than in design-bid-build. Here's why.

Why Design-Build Engineering Raises the Stakes

Design-build engineering is a project delivery method in which a single entity holds one contract for both design and construction³— eliminating the separate-contract structure of design-bid-build. That single contract is what makes the design review more consequential: a missed issue surfaces in construction, not in a redesign cycle.

In design-build engineering, the design review is the last cheap place to find a problem. Every dollar spent fixing an issue at the design phase saves multiples downstream. The methodology is also winning. FMI Corporation, in partnership with DBIA⁴, projects design-build will represent over 47% of US construction spending by 2028, totaling more than $1.1 trillion across the 2024–2028 forecast period. More projects, same broken meetings.

Over-the-Shoulder (OTS) reviews— a project administration tool used in design-build and CM-GC delivery to let the owner observe and provide feedback on work in progress⁵— already exist on most D-B projects as the informal version of what a rubric formalizes. The framework anchor is DBIA's Design-Build Done Right® Universal Best Practices, updated in 2023⁶ with four Guiding Principles on top of 10 best practices. A rubric is compatible with that framework, not a replacement for it.

If the design review is the cheap place to find a problem, the question is what instrument actually finds it. Most teams reach for a checklist. That's not enough.

The Rubric Replaces the Checklist (Not the Other Way Around)

A checklist verifies that something is present. A rubric scores how good it is, against defined anchors and weighted by criticality. Most AEC firms run their design reviews on checklists— verifying presence— and then wonder why the meeting still has to do all the work.

A scoring rubric, per ASEE's Engineering Design Process Portfolio Scoring Rubric research⁷, includes dimensions or criteria on which performance is rated, definitions illustrating measured attributes, and a rating scale for each dimension. In other words: criteria, anchors, scale. The same item looks different on each instrument.

A checklist confirms completeness. A rubric forces judgment. The two are not enemies. Discipline-specific design review checklists from the Whole Building Design Guide (WBDG) covering Architectural, Structural, MEP, Electrical, and other domains⁸ are excellent inputs to a rubric— they tell you what to score. WSDOT's constructability review process⁹ uses stage-scaled checklists for the same reason. Checklist is the necessary precursor. Rubric is the next step.

The "why this matters" is plain. Live discussion in a meeting should be reserved for the few criteria that scored low. If everything is checked off, you don't know which items deserve airtime. And that's how you get a two-hour meeting that reviews nothing.

Now we can build the rubric. Four pieces: criteria, scale, weights, and anchors.

Anatomy of a Working Design Review Rubric

A working design review rubric has four parts: criteria (what gets scored), a scale (how it's scored), weights (how much each criterion counts), and anchors (what each score actually means). Get any one of these wrong and the rubric becomes theater.

1. Criteria. What you score. Section 5 gives the recommended list. Nine is enough; twenty is a checklist in disguise.
2. Scale. Use a 5-point scale (1 = does not meet; 3 = meets; 5 = exceeds). A 10-point scale invents false precision and turns the score into haggling. Five forces commitment.
3. Weights. Each criterion has a weight, summing to 100%. Safety, code compliance, and constructability typically carry 2–3x the weight of design intent items. This is where leadership encodes priorities. Safety doesn't get out-voted by aesthetics if safety is weighted three times higher.
4. Anchors. Each score (1–5) has a written definition. Without anchors, every score collapses to a 4 and you've learned nothing.

A worked anchor for one criterion looks like this:

Constructability— Score of 3 (Meets): Design is buildable as drawn. Standard means and methods apply. Independent constructability reviewer (if used) found no major flags. No clashes flagged in the BIM coordination model at the current level of detail.

A 5 on that same criterion would mean the constructability reviewer verified the design and identified value-engineering opportunities. A 1 means the design as drawn is not buildable without redesign. Without that level of specificity, the score is sentiment. With it, the score is evidence.

A rubric's weights are where leadership encodes priorities. Without anchors, every score is a 4 out of 5— and you've learned nothing.

Now the criteria. Here are nine that earn their place on a design-build rubric.

Recommended Criteria for a Design-Build Engineering Review Rubric

Nine criteria belong on a design-build engineering review rubric: code/standards compliance, constructability, MEP/BIM coordination, schedule fit, budget fit, owner program alignment, safety, value engineering, and sustainability. Each gets a weight and an anchor.

Weights are starting points, not industry standard. A hospital project weights life safety higher than a parking structure. An owner-occupied corporate campus weights sustainability higher than a logistics warehouse. Adjust to project type. WBDG's discipline checklists feed the technical anchors⁸; DBIA Best Practices anchor the methodology⁶. Constructability and MEP coordination together should account for at least 30% of total weight on a design-build rubric— these are where surprises become rework.

If "owner program alignment" isn't on your rubric, you're scoring the design without scoring whether it's the right design. And constructability anchors should scale across stages: a 3 at schematic isn't the same depth as a 3 at IFC, which is exactly how WSDOT structures its constructability reviews⁹. Same criterion, different anchor at each stage.

The criteria are the easy part. The hard part is the protocol that makes the rubric actually shorten the meeting.

The Protocol — Pre-Meeting Scoring, 30-Minute Reconciliation, Action Register

The rubric only shortens the meeting if scoring happens before the meeting. The protocol has three steps: discipline leads score independently 48 hours before the meeting; the meeting itself reconciles only contested criteria; an action register captures every score below 3.

1. Independent pre-scoring (48 hours pre-meeting). Discipline leads— architectural, structural, MEP, civil— score the rubric independently against their domain. No collaboration during scoring. Submitted to the design-build PM 48 hours before the meeting. Independent scoring is what surfaces real disagreement. Group scoring surfaces consensus, which is a different signal.
2. Aggregation and reconciliation meeting (24 hours pre-meeting; 30-minute target). PM circulates aggregated scores 24 hours before the meeting. Meeting agenda is mechanical: the only criteria discussed are those where scores diverged across reviewers, plus any criterion scored below 3. When 80% of the rubric scores "meets" before the meeting starts, only the contested 20% needs live discussion. That's how a two-hour review becomes a thirty-minute reconciliation.
3. Action register (live, during meeting). Every sub-3 score generates an action item: owner, due date, the specific anchor language being violated. Linked to the design package directly, not the meeting minutes. Minutes get lost; the design package follows the project.

Time-box target: 30 minutes for typical phase reviews. 60 minutes for complex stage gates (DD-to-CD transition, major program changes). Anything longer is a sign the rubric isn't being used.

The meeting is for arguments. The scoring is for evidence. Bliley¹ says it plainly: ruthlessly insist that everyone comes prepared and ready to jump into the difficult tasks of improving the design. Pre-prep that isn't paid for doesn't happen, which is why the rubric needs to be a billable deliverable scoped into the design phase. More on that in the watch-outs section.

Who fills out what is the next problem. RACI is the right tool.

Roles — Who Scores, Who Decides, Who's Informed

Discipline leads score independently. The design-build PM is accountable for reconciliation. The owner is consulted, not voting. Subcontractors are informed of outcomes that affect their scope.

RACI¹⁰ gives every task one Responsible doer and exactly one Accountable owner. Applied to the rubric:

Every criterion has exactly one accountable owner. If the constructability score is contested and no one is accountable, the score is theater. The owner consults— the owner does not score. Owner-as-scorer turns the rubric into a negotiation, which is the dysfunction the rubric exists to fix.

Independent constructability reviewers earn their seat in the Consulted column for one reason: you can't read the label from inside the bottle. WSDOT⁹ uses external review at key stages — non-design-team reviewers often catch what internal teams overlook. A reviewer outside the design team scores the truth that internal teams won't.

The rubric doesn't replace constructability and BIM coordination— it connects them. Here's how.

How the Rubric Connects to Constructability, OTS, and BIM Coordination

The rubric is the formal scoring layer above the existing review stack. Constructability reviews feed the constructability score; BIM clash detection feeds the MEP coordination score; Over-the-Shoulder reviews are the informal version that the rubric formalizes at the stage gate.

• Constructability review (per WSDOT⁹) feeds the Constructability criterion. The constructability checklist is the input; the rubric is the score. Stage-scaled checklists become stage-scaled anchors.
• BIM clash detection is downstream verification of design coordination. If 932 clashes go unresolved at design phase, the rubric should have caught the coordination problem before the model did. The rubric is the leading indicator; clash detection is the lagging one.
• OTS reviews⁵ are the informal D-B-specific tool. The rubric is the formal complement at the stage gate. Same engineers, same project, different cadence.

The proof point for what structured coordination is worth comes from one DBIA-published case study¹¹. On a $230 million design-build food manufacturing project, $200,000 of BIM coordination labor delivered $2.22 million in rework cost avoidance and $542,000 in schedule savings— a net $2.55 million benefit, roughly 10x ROI on the coordination investment. 932 clashes were resolved out of 2,000 raised. Top categories: structural steel, conveyor equipment, electrical power.

One DBIA-published $230M project: $200K of BIM coordination labor → $2.22M rework avoided + $542K schedule savings = $2.55M net (10x ROI).

Critical attribution: this is one named project, not an industry average. At the macro level, FMI— reported via Trimble¹²— estimates more than $177 billion is lost each year in US construction to inefficiencies including rework, search time, and communication breakdowns, with construction professionals spending 14+ hours per week on non-core tasks. The case study is what one structured project did. The macro number is what the industry leaves on the table. This kind of impact is the case for measuring AI success across operational changes the same way you'd measure any process change— with leading and lagging indicators paired.

Three things will kill a good rubric. Here's how to handle each.

Watch-Outs — How the Rubric Gets Gamed (and How to Stop It)

Three failure modes will kill this rubric: everyone scores 4/5 to avoid conflict; pre-meeting prep becomes unbillable overtime; high rubric scores get treated as guarantees instead of leading indicators. Each has a fix.

• Failure 1: Score compression at 4. If every criterion scores a 4, the rubric isn't measuring quality— it's measuring conflict aversion. Fix: require written rationale for any score of 4 or 5 on a contested criterion. Make the anchors specific (e.g., "5 = independent reviewer verified"). Optional: anonymous initial scoring to reduce social pressure.
• Failure 2: Pre-prep as unbillable overtime. Pre-prep that isn't paid for doesn't happen. Fix: the rubric is a billable deliverable, scoped into the design phase contract. The hours that move from meeting to pre-prep are the same hours, just spent better. This is also where building a culture that adopts new instruments matters— a rubric the firm doesn't pay for is a rubric the firm doesn't run.
• Failure 3: Rubric score treated as causation. A high score predicts fewer RFIs and lower change-order rates. It does not guarantee them. Fix: position the rubric as a leading indicator and pair it with downstream measurement— RFI count, change-order rate, rework hours. Use the downstream data to recalibrate weights and anchors quarterly.

The rubric is also where AI becomes useful— not as the reviewer, but as the second set of eyes on the data.

Where AI Augments the Rubric

AI doesn't replace the engineer scoring the rubric. It pre-fills the evidence the engineer needs to score it. Three places this works today: clash detection summaries, code compliance cross-checks, and document-difference reports between design iterations.

• Clash summarization. AI can read a BIM clash report and surface the high-priority subset from the 2,000 raised— the same filtering work that, on the DBIA case¹¹, narrowed 2,000 detected clashes to 932 resolved by the human team. The MEP lead still scores; AI just makes the evidence reviewable in 20 minutes instead of 4 hours.
• Code compliance cross-check. AI can compare a design package against IBC and known jurisdictional amendments and flag potential issues for the engineer to verify. This is where how founders should evaluate AI investments matters— the engineer scores, AI does the discovery, and the firm decides which workflows are worth automating.
• Iteration diffs. AI can summarize what changed between design package versions, making rubric re-scoring at later stages faster. At schematic-to-DD transitions, this turns hours of side-by-side comparison into a structured change list.

Construction productivity improved only 10% between 2000 and 2022, per McKinsey¹³— a $1.6 trillion gap with other sectors. The rubric closes a small piece of it. AI augmentation closes a slightly larger piece. Both are intellectual augmentation, not substitution. The engineer still scores. AI just makes sure the engineer is scoring against the right facts.

Wherever you start, start with one rubric on one project. Then watch what the next meeting looks like.

Take It and Run — A Working Rubric Template

The rubric template above (Section 5's criteria + Section 4's scale + Section 7's roles) is the minimum viable instrument. Pick one project, run one rubric, see what the next meeting looks like.

• Pilot. One project. One stage gate (DD or CD review is best). Run the protocol exactly as written: 48-hour independent scoring, 24-hour aggregation, 30-minute reconciliation.
• Revise. After the pilot, adjust weights and anchors based on what scored compressed (everyone at 4) versus what split. Quarterly review or after each major project type.
• Standardize. Once two pilots run cleanly, standardize the rubric across the firm and scope it as a billable deliverable on every D-B engagement.

The first rubric is rough. The third one is faster than the meeting it replaces. Better thinking before the meeting is what shortens it. The rubric is just the instrument that forces the thinking.

If your firm is integrating design review rubrics with AI-augmented workflows— clash filtering, code cross-checks, document iteration diffs— an AI implementation partner who works with AEC firms can help architect those systems without locking you into a vendor.

Dan Cumberland Labs' work with founder-led firms sits at the intersection of operational instruments like this rubric and the AI workflows that pre-fill the evidence behind them. The rubric is the instrument. The workflow around it is where the leverage lives.

FAQ

How is a design review rubric different from a constructability checklist?

A constructability checklist verifies presence— is the structural calc included, is the MEP layout coordinated, is the egress path drawn. A design review rubric scores how good those items are, against weighted criteria with anchored definitions per score level⁷. WBDG's discipline checklists⁸ feed the rubric as inputs. The checklist tells you what to look at; the rubric tells you how it stacks up. Most AEC firms have the checklist already. Adding the rubric layer is the next step, not a replacement.

Can the same rubric work across schematic, DD, CD, and IFC stages?

Yes, with stage-scaled anchors. The criteria stay the same across stages; the anchor definitions for each score change. WSDOT's constructability review process⁹ follows this pattern— same checklist categories, different depth at each design phase. A score of 3 at schematic means the design is conceptually buildable; a score of 3 at IFC means the design is buildable as drawn with verified means and methods. Pilot the rubric at one stage first, then scale.

How often should the rubric criteria and weights be revised?

Quarterly review at minimum, plus a post-project debrief after each major project type. Track downstream signals— RFI counts, change-order rates, rework hours per project— and use that data to recalibrate weights. If a project with a high coordination score still produced 200+ RFIs in construction, the rubric's anchors for coordination need tightening. Weights are starting points, not industry standard. The firm that runs the rubric for two years and never adjusts it is running a snapshot of last year's priorities.

References

1. Bliley Technologies, "5 Simple Steps for Truly Effective Design Reviews" (2020) — https://blog.bliley.com/5-simple-steps-for-truly-effective-design-reviews
2. Atlassian (covered by Fortune), "Meetings are a productivity killer—and 3 in every 4 are totally ineffective, according to a new wide-ranging study" (2024) — https://fortune.com/2024/03/21/meetings-productivity-ineffective-atlassian-report/
3. Design-Build Institute of America, "What Is Design-Build?" (2024) — https://dbia.org/what-is-design-build/
4. FMI Corporation in partnership with DBIA, "New FMI Report Highlights Design-Build's Role in Tackling Industry Challenges and Driving Growth (2024 Design-Build Utilization Study)" (2025) — https://dbia.org/blog/new-fmi-report-2024/
5. Arizona State University, Alliance for Construction Excellence, "Session 5 – Tool: Over the Shoulders Review – D-B and CM-GC Contract Administration" (2020) — https://acm.engineering.asu.edu/session-5-tool-over-the-shoulders-review/
6. Design-Build Institute of America, "Updated Design-Build Done Right® Universal Best Practices Available Now" (2023) — https://dbia.org/blog/updated-design-build-done-right-universal-best-practices-available-now/
7. American Society for Engineering Education (ASEE), "The Engineering Design Process Portfolio Scoring Rubric (EDPPSR) – Initial Validity and Reliability" (2014) — https://peer.asee.org/the-engineering-design-process-portfolio-scoring-rubric-edppsr-initial-validity-and-reliability-fundamental
8. Whole Building Design Guide / National Institute of Building Sciences, "Design Review Checklists and Operation Checklist" (2024) — https://www.wbdg.org/ffc/va/design-review-checklists
9. Washington State Department of Transportation (WSDOT), "Project Delivery Constructability Review (Implementation of the Constructability Review Process)" (2014) — https://wsdot.wa.gov/publications/fulltext/ProjectMgmt/Website/Implementation-of-the-Constructability-Review-Process.pdf
10. Asana, "RACI Charts: The Ultimate Guide, with Examples" (2025) — https://asana.com/resources/raci-chart
11. Design-Build Institute of America, "The True Value of Clash Detection: A Detailed Return on Investment (ROI) Case Study" (2024) — https://dbia.org/blog/the-true-value-of-clash-detection-a-detailed-return-on-investment-roi-case-study/
12. FMI Corporation (reported via Trimble), "How poor design drives $177B in construction rework" (2025) — https://www.trimble.com/blog/construction/en-US/article/collaborative-design-sketchup-cut-construction-rework-costs
13. McKinsey & Company, "Improving construction productivity is the new imperative" (2024) — https://www.mckinsey.com/capabilities/operations/our-insights/delivering-on-construction-productivity-is-no-longer-optional