Analytics Metrics Glossary
Every metric BayWise reports, what it counts, and what to do when it moves.
→ For how to use the screen, see Understand analytics.
Overview
| Metric | Definition | Healthy |
|---|---|---|
| Total Jobs | Jobs in the selected period | — |
| Completed | Jobs delivered in the period | — |
| Revenue | Billed value of completed work | — |
| Cost | Labour cost of completed work | — |
| Gross Profit | Revenue minus labour cost | 50%+ margin per job |
| Avg Cycle Time | Average time from a job starting to being delivered | Depends on your work mix |
Daily Completions Trend — jobs delivered per day. Consistent bars mean steady throughput. Gaps signal blocked bays, missing parts or understaffing.
Bay utilisation
| Metric | Definition | Healthy |
|---|---|---|
| Average Bay Utilisation | Percentage of operating hours each bay had active work | 70–85% |
| Utilisation Trend | Day-by-day average across all bays | Stable or rising |
| Peak | The busiest point in the period | — |
| Idle bays | Bays with no active work in the period | 0 |
Below 60% is capacity you are paying for and not using. Above 90% leaves no room to absorb a bad morning, and is usually accompanied by rising promise misses.
A bay with a holding or delivery role does no work by design, and low utilisation on it means nothing. Read this metric across working bays only.
Technicians
| Metric | Definition | Healthy |
|---|---|---|
| Productivity Ranking | Total billable hours worked per technician | Gaps explainable by role |
| Revenue per technician | Revenue billed against their work | — |
| Profit per technician | Revenue minus their labour cost | Positive for everyone |
| Eff% | Billable time divided by clocked-in time | 70%+ |
Large gaps between ranks usually mean skill mismatches or uneven allocation rather than uneven effort — and the fix is normally in how skills are recorded, not in a conversation with the person at the bottom.
Clocked hours are read from attendance. If attendance cannot be loaded, cost and efficiency are not shown rather than being estimated.
Efficiency and estimates
| Metric | Definition | Healthy |
|---|---|---|
| Estimation Accuracy | Actual completion time as a percentage of estimated time | 95–105% |
| Accuracy Trend | Whether accuracy is moving toward 100% over time | Trending to 100% |
| Per-Service Accuracy | Accuracy broken down by service | Clustered near 100% |
Read it as: 100% is exact. Below 100% means faster than estimated. Above 100% means slower.
This is the most directly actionable screen in BayWise. A service consistently above 100% has an optimistic catalogue duration, and every job of that type is quietly manufacturing promise risk. Raise the duration.
A single day’s spike is one difficult job, not a systemic problem. Two data points are needed per service before accuracy is shown at all.
Delays and promises
| Metric | Definition | Healthy |
|---|---|---|
| Promise Adherence | Percentage of jobs delivered on or before the promise | 85%+ |
| On time / Late | Counts behind the percentage | — |
| Avg delay | Average lateness across late deliveries | — |
| Promise changes | How often promises were moved | Low and deliberate |
| Late Deliveries | The individual jobs that missed, with vehicle, promised time, delivered time and delay | — |
| Currently Overdue Steps | Active stages past their estimated end time right now | 0 |
Chronic promise misses almost always trace back to one of two causes: optimistic catalogue durations, or parts. The efficiency screen tells you which.
A high count of promise changes is not automatically bad — a promise moved with the customer’s agreement is good practice. It is bad when it is being used to make the adherence number look better than the customer’s experience was.
Service mix
Service Frequency — how often each service was performed in the period. Useful for deciding what to specialise in, what to price differently, and which catalogue durations are worth the effort of getting exactly right.
Periods
Analytics can be filtered by day, week, month, or a custom range. Every metric above respects the selected period.
Trend charts are calculated across working days only, so a bank holiday does not appear as a collapse in throughput.