The clash matrix — the grid of discipline-pair comparisons that forms the backbone of a Navisworks coordination run — is a useful organizational tool that most teams manage poorly. The result is coordination meetings dominated by administrative overhead: sorting through an 800-row clash list, arguing about which items are real, and losing track of what was resolved three rounds ago. None of that overhead makes the building better.
The practices that follow are not software tricks. They are process decisions — most of which cost nothing to implement and produce immediate improvement in the signal-to-noise ratio of your coordination data.
Start with a Rule Set Audit Before Every Project Phase
Most firms inherit Navisworks rule sets from previous projects and apply them without review. The rule set that worked for a 6-story office building may not be appropriate for a healthcare facility with different clearance requirements or a parking structure with different structural geometry. Running a rule set without auditing it for the current project type is the fastest way to generate a clash list that is half irrelevant.
A rule set audit at the start of each project phase (SD, DD, CD) takes 30 minutes and covers:
- Tolerance settings: hard clash tolerance should be 0.00 inches for most structural-MEP pairs. Clearance clash tolerances should reflect the actual maintenance access requirements for the project's mechanical systems — typically 18–24 inches for major air handling equipment, 12 inches for duct runs. Do not copy tolerances from a previous project without verifying they match this project's specifications.
- Included categories: are you including insulation geometry in MEP clashes? For coordination against structural, insulated duct dimensions (not bare duct) should be the clash surface. Running bare duct dimensions against structure and then discovering that the duct plus insulation doesn't clear is a coordination failure that happens regularly.
- Discipline pair coverage: confirm that every relevant discipline pairing is in the matrix. Common gaps: civil/site vs. foundation, fire protection vs. electrical, architectural millwork vs. MEP. The last one generates surprisingly many field RFIs on projects with dense millwork in mechanical rooms.
Classify Clashes at the Source, Not in the Meeting
The most common coordination meeting failure is bringing an unclassified clash list to the room. When the coordinator distributes a 600-row Navisworks report 30 minutes before the meeting, every participant is doing triage in real time instead of making decisions. The result is a meeting where 70% of the time is spent on classification and 30% on resolution — when it should be the inverse.
Classification before the meeting requires the BIM Coordinator to assign at minimum three fields to every clash in the list before distribution:
- Priority: Critical (blocks critical-path design decisions), High (requires resolution before next coordination round), Medium (should be resolved before CD issue), Low (housekeeping, resolve when convenient)
- Responsible discipline: who owns the design decision that needs to change? Not "both parties need to discuss" — a specific discipline lead is assigned. If ownership is genuinely unclear, that ambiguity is itself the first agenda item for that issue.
- Status: New, In Progress (assigned, solution proposed), Resolved (model updated), Closed (confirmed resolved in updated model). Never let a clash sit in "Resolved" without confirming the model update.
Pre-classified clash lists cut meeting time roughly in half. The agenda can then focus exclusively on Critical and High priority items, with Medium and Low acknowledged in bulk and assigned.
The Duplicate Problem and How to Prevent It
Navisworks' clash detection reports every geometric intersection as a separate clash. A single duct run that is incorrectly routed through a structural beam may generate 15–30 individual clash entries — one for each pair of geometry faces that intersect. To a human eye, this is one routing problem. To the raw clash list, it is 30 line items.
The grouping step — collapsing spatially related clashes into single coordination issues — must happen before the clash list reaches the coordination meeting. Manual grouping in Navisworks uses clash grouping by selection set or by spatial proximity. The practical approach: after running the clash detection, select all clashes within a defined spatial zone (say, 10 feet in each direction from a central point) and group them. A 30-item duct-beam conflict in a mechanical room corridor becomes one grouped issue with 30 associated sub-clashes.
Grouped issues should be assigned a single Clash ID and tracked as one coordination item. When the duct routing is corrected, all 30 sub-clashes resolve simultaneously. Tracking them individually creates administrative overhead proportional to the grouping multiplier — and on complex projects, that multiplier can run 10x or higher.
Keeping Resolved Items Out of the Active List
A coordination matrix that accumulates every clash from every run without pruning becomes unmanageable by Round 4. The discipline leads stop reading the full report because they know most of it is stale. They scan for their discipline's new items, miss something in the middle, and coordination breaks down.
The protocol for keeping the active list clean:
- After each model update from a discipline, re-run clash detection for only that discipline's pairs
- Compare new results against previous round: clashes that no longer appear in the new run are candidates for closure
- Before closing a clash, confirm the updated model was submitted and the coordination issue database reflects the resolution
- Archive closed items in a separate tab or section — accessible for reference but not in the active working view
The active coordination matrix should contain only open and in-progress items. When a principal-in-charge asks "how many open coordination issues do we have?", the answer should come from the active list without having to mentally subtract resolved and closed items.
The Severity Calibration Problem
Severity scoring in many firms defaults to geometric: a larger overlap volume equals higher severity. This is a proxy metric that is often wrong. A 2-inch overlap between an MEP conduit and a concrete wall penetration in a non-structural location is a lower-severity issue than a 0.5-inch clearance violation between a fire damper and a structural column in a rated fire wall assembly. The latter is a fire/life safety compliance issue that requires code analysis, not just a routing adjustment. The former is a minor penetration sleeve coordination item.
Severity should be based on consequences, not geometry. The five factors that drive consequence-based severity scoring:
- Whether the resolution requires structural changes (vs. MEP rerouting, which is cheaper)
- Whether the conflict zone is in a rated assembly (fire wall, smoke barrier)
- Whether the conflict affects a critical-path structural element (moment frame, lateral system)
- Whether the conflict is in a zone with constrained access during construction (below-grade mechanical rooms, within concrete pour zones)
- LOD status of both conflicting elements — lower LOD means lower confidence in severity
We are not saying geometric overlap is irrelevant — it is a useful first-pass signal. A 12-inch hard clash between a W18 beam flange and a 30-inch round duct is almost certainly a high-priority routing conflict regardless of context. The calibration matters for the medium-severity range, where geometric proximity alone is an unreliable guide to priority.
Discipline-Specific View Configurations
One practical time-saving measure that few teams implement consistently: save named Navisworks search sets and viewpoints for each discipline lead before the coordination meeting. The structural engineer should be able to open the federated model and immediately activate a view that shows only their issues — not every clash in the project, filtered to their discipline pairs, with the clash marker centered in view for each item.
This requires 45–60 minutes of setup before each coordination round but eliminates the meeting time typically spent with a coordinator screen-sharing and navigating the model while discipline leads wait to see their issues. Each discipline lead can review their items independently before the meeting, arrive with a position on each, and the meeting moves directly to decision-making.
The combination of pre-classification, pre-grouping, pre-distribution, and discipline-specific views produces a different kind of coordination meeting. It is shorter. It is less contentious, because participants are not surprised by their issues. It produces more decisions per meeting-hour. And it generates a cleaner coordination register — because items that were pre-reviewed are resolved more definitively in the meeting than items discovered in real time.