BIM coordination checklist: 12 checks before every clash review
Twelve practical checks for coordinates, model versions, clash tests, tolerances and issue tracking, so your next clash review finds real problems instead of noise.
Most bad clash reviews go wrong before anyone opens the clash report. The models were from different weeks, the coordinates were off by 300 mm, and “everything against everything” produced 6,000 results nobody will read.
Run through these twelve checks before the next review. They take less time than the meeting they save.
Before you run anything
1. Everyone shares the same coordinates. Check that the survey point and project base point agree across every model, or that everyone links by shared coordinates. A whole model shifted by a fixed distance is the classic symptom.
2. Levels and grids match. Overlay the levels and grids from each discipline. A slab 50 mm off between the architectural and structural models will generate hundreds of fake MEP clashes.
3. Every model is the agreed version. Note the date or revision of each model you load. Coordinating last week’s MEP against this week’s structure is a waste of everyone’s afternoon.
4. Naming and units follow the BIM execution plan. Files, levels, systems and worksets named the agreed way make clash results readable and filterable.
5. Only the right content is exported. Leave out unloaded links, temporary elements, design options nobody chose and placeholder families. Each one adds noise.
6. Each model is clean on its own. Review warnings, delete duplicate and overlapping elements, and check that MEP systems are connected. A discipline should not send its internal problems to coordination.
Setting up the tests
7. Test by system pairs, not everything against everything. Ducts against structure, pipes against structure, ducts against pipes, cable trays against ducts. Smaller tests produce results that one person can own.
8. Use tolerances that fit the trade. A hard clash tolerance of a few millimetres hides modelling noise without hiding real problems. Add clearance tests where they matter: insulation thickness, maintenance access in front of equipment, space for sprinkler heads.
9. Group results into issues. A duct running through ten beams is one issue with ten clashes, not ten issues. Group by element, zone or level so the list matches the decisions you need to make.
Running the review
10. Every issue gets an owner, a due date and a status. Use BCF or your common data environment, so issues travel back into the authoring tools instead of living in a slide deck.
11. Walk the agenda by zone and priority. Start with the ceilings that close first on site, or the risers that block everything else. Ten decided issues beat forty discussed ones.
12. Close the loop. Re-run the same tests after fixes, with the same settings, and track the numbers from review to review. A count that stops falling tells you something is wrong with the process, not just the model.
Keep it repeatable
Save the test set, the tolerances and the agenda template, and reuse them on the next project. Coordination gets faster when it stops being reinvented every time.
The BIM coordination, made practical course covers this whole loop in more depth, from planning exchanges to running meetings that close issues.