top of page

HVAC RFI Dashboards: Separate Reply Rate from Verified Implementation

Oct 15
5 min read
Original blue and lime Nexora cover showing four progressively smaller RFI counts: twelve issued, ten replied, eight usable and five evidence-complete.

An HVAC coordination report shows that 83% of RFIs have received a reply. That sounds encouraging, but the same register may contain incomplete answers and drawings still awaiting revision. A useful dashboard separates receipt, usable decisions and verified implementation. It also explains its denominator so a change in the percentage can be interpreted correctly.


1. Define the cohort and the evidence cutoff


Start with a named group of RFIs and one reporting cutoff. A floor, equipment package or project phase can provide a useful cohort. Decide whether the report includes all issued RFIs or only active, applicable items. Record withdrawn, duplicate and not-applicable entries separately. Do not remove a difficult question merely to improve the completion percentage.


Keep the formal RFI status required by the project procedure. The stages below are additional management measures, not replacement contractual statuses. Procore’s closure documentation allows an RFI to be closed without a response. A count of records marked closed therefore needs interpretation before it is used as evidence that technical work is complete.


Freeze the reporting snapshot. A response received after the cutoff belongs in the next update, even if the register is edited before the report is circulated. Keep the source response and evidence timestamps so a reviewer can reproduce the figures without relying on today’s live status.


2. Create nested checkpoints with explicit rules


  • Reply received: at least one response is present by the cutoff. An acknowledgement can satisfy receipt, but not the next checkpoint.

  • Usable answer: the project’s authorized response resolves the question, identifies the controlling references and states any conditions needed for the affected work.

  • Implementation evidence complete: all applicable downstream actions for the current answer have been checked, or a justified no-action decision has been reviewed and recorded.


Every item reaching a later checkpoint must have passed the earlier ones. These are cumulative counts, so they must not be added together. An RFI with checked implementation evidence appears in all three totals. Keep the rules beside the dashboard and use the same interpretation across packages.


Procore’s official-response guidance distinguishes an authorized answer within a longer discussion and advises seeking a clear answer when the thread remains ambiguous. Do not turn several conflicting comments into a usable decision merely to advance a dashboard stage.


3. Work through a twelve-RFI snapshot


Assume an original teaching cohort of twelve active, applicable RFIs, with no withdrawn or duplicate entries. Ten have replies; eight of those ten have usable authorized answers; five of those eight have complete checked implementation evidence. All counts refer to the same cutoff and current response revisions.


  • Reply-received rate = 10/12 × 100 = 83.3%.

  • Usable-answer rate = 8/12 × 100 = 66.7%.

  • Implementation-evidence rate = 5/12 × 100 = 41.7%.


Now convert the cumulative stages into mutually exclusive action groups. Unanswered = 12 − 10 = 2 RFIs. Replies needing clarification = 10 − 8 = 2. Usable answers awaiting implementation evidence = 8 − 5 = 3. Evidence complete = 5. The reconciliation is 2 + 2 + 3 + 5 = 12.


This partition makes the next action visible. The two unanswered items need response follow-up; the two incomplete answers need clarification; the three implementation items need their downstream owners. Adding 10 + 8 + 5 would count some RFIs several times and produce a meaningless total of 23. State the raw counts alongside the rounded percentages.


Original RFI dashboard diagram showing twelve total items split into two unanswered, two needing clarification, three awaiting implementation evidence and five evidence-complete, with cumulative rates of 83.3, 66.7 and 41.7 percent.

Original fixed-cohort example at one cutoff. Cumulative stage counts overlap; mutually exclusive action groups sum to twelve. Rates do not measure physical progress or risk.


4. Link the last checkpoint to a revision-based matrix


For example, a hypothetical RFI-042 response R1 resolves a chilled-water branch routing question through an authorized sketch. Its implementation matrix links four outputs to that exact response: coordinated drawing M-214 Rev D, fabrication list F-042 Rev 2, reviewed support layout S-021 Rev C and field work pack WP-17 with its revised drawing manifest. The routing decision itself is assumed to have received the required technical review.


For each output, record an owner, required revision, evidence reference, reviewer and check date. Suppose the coordinated drawing and fabrication list are checked, while the support layout is under review and the field pack still contains the previous drawing. This RFI remains in the usable-answer group awaiting implementation evidence. Two completed outputs out of four do not make half an RFI complete in the binary dashboard.


A separate action-completion metric can be useful, but label its denominator as actions. Adding many minor actions to one RFI can change an action-based percentage without changing any engineering consequence. An empty matrix row is unresolved; a no-action conclusion needs a stated reason and reviewer.


Procore’s drawing-link documentation supports connecting RFIs and other records to drawing locations. Such links make evidence easier to retrieve. They do not prove that a fabrication package, field drawing or installed arrangement matches the current answer; that remains a defined review activity.


5. Explain percentage changes before judging performance


Consider a separate next-period scenario: the original twelve RFIs remain unchanged, but one new unanswered RFI is added. The counts become thirteen issued, ten replied, eight usable and five evidence-complete. The implementation-evidence rate falls from 41.7% to 5/13 × 100 = 38.5%, even though no previously completed item regressed. The denominator expanded.


Show opening items, new items, withdrawals, completions and reopened items alongside the closing balance. To compare like for like, retain a fixed-cohort view for the original twelve and a separate current-register view for all thirteen. A raw trend line without those movements can reward quiet periods and penalize a team for properly recording new questions.


6. Invalidate stale evidence when an answer changes


Return to the original twelve-RFI snapshot for a different scenario. If response R2 supersedes R1 on one of the five evidence-complete RFIs, assess which outputs relied on R1. Suppose R2 is already a usable authorized answer but affected outputs require rechecking. The evidence-complete count becomes four, giving 4/12 × 100 = 33.3%; the usable-answer count remains eight. The implementation queue rises from three to four.


Preserve the previous evidence and reopen or revise the linked implementation actions through the project procedure. A checked status belongs to a specific response revision and output set. It must not automatically carry forward to a changed instruction. If the new answer is itself unclear, reassess the usable-answer checkpoint too.


These counts measure workflow evidence, not physical progress, cost exposure, contractual entitlement or schedule delay. One unresolved RFI may have a much greater consequence than several completed ones. Keep the affected activity, required decision and consequence visible next to the numerical summary.


Nexora’s paid HVAC RFI Pack provides a register, form, tracker, dashboard and HVAC question library for organizing technical questions and follow-up. Use its record references as a starting point for a project-specific evidence trail. The three-stage metrics and implementation matrix described here are an original supplementary method, not a claim that the pack automatically provides or verifies them.



Comments

Rated 0 out of 5 stars.
No ratings yet

Add a rating
bottom of page