Reps grouped by department. Equipment-mix target 33–40%. Add-Ons = stickers, signs, mounting, plexiglass sold. Big = invoices over $5,000.
By Service Type
Ranked by total invoice.
Time Utilization
The working day is 7:00 AM to 3:30 PM, extended earlier to a tech’s first check-in when they start early and later to their last check-out when they run late. On-job is measured from check-in/out; idle & parked come from vehicle telematics clipped to that window (overnight/weekend yard time is excluded). Driving is inferred as the remainder of the window, so the four add up to the working day and utilization can’t exceed 100%. A route with no movement and no jobs is treated as not working that day. This applies to data collected after deploy; earlier dates can’t be re-clipped.
How the buckets are accounted
Each route’s working day is split into six buckets that add up to the shift window:
On Job — time on site, measured from invoice check-in/check-out. Productive.
Driving — measured from GPS movement (speed above 5 mph). Productive.
Depot/Prep — stationary at the warehouse (loading & prep), measured by a 500 ft geofence. Productive.
Idle — engine on but not moving, out in the field.
Parked — engine off and stationary, out in the field.
Unaccounted — the leftover after the five above (GPS gaps, brief blips). A data-quality signal, not real work.
Utilization = (On Job + Driving + Depot) ÷ shift window. Occupancy = On Job ÷ available time (window minus a 1-hour lunch). The window opens at the earlier of 7:00 AM or first check-in, and closes at the latest of 3:30 PM, last check-out, or the drive back to the warehouse. Driving and Depot are measured from GPS collected after this feature was deployed; earlier days fall back to an estimated Driving figure, show no Depot, and no Unaccounted.
Line Items
Capacity & Pricing Signals · decision support — suggestions, a person decides
Quartiles of measured on-site time (service work is right-skewed, so mean alone misleads) turned into a realistic capacity range and pricing/predictability flags. Capacity is against a 420-min day and excludes travel, so real capacity is a bit lower. Buckets under 30 samples are still collecting. Revenue figures fill in as history is re-harvested.
What these numbers mean
Think of every job of one type lined up fastest to slowest.
Median — the job right in the middle of that line. Half of jobs are faster, half are slower. This is the typical job — quote and schedule against it. One or two huge jobs don’t move it.
Mean (average) — add every job’s time and divide by how many. If the mean is a lot higher than the median, a few very long jobs are pulling the average up.
Q1 (25%) — a quarter of the way down the line. A quarter of jobs finish faster than this. Your “clean, quick job” time.
Q3 (75%) — three-quarters of the way down. A quarter of jobs take longer than this. Plan a careful day around Q3, or you’ll run late on about 1 in 4 jobs.
Min / Max — the fastest and slowest job ever recorded for this type. Max is worth a look — it’s often a hard site or a data glitch.
How to read the box chart
The blue box covers the middle half of jobs (Q1 to Q3). The dark line inside it is the typical (median) job. The thin line stretching out each side reaches the fastest and slowest. A short box = jobs take about the same time every time. A long box = times are all over the place.
Predictability — how much the box widens
Tight — jobs take about the same time every time. Trust the median, schedule them back-to-back.
Moderate — some job-to-job variation. Leave a little buffer in the schedule.
Wide — times are unpredictable. Don’t pack the day tight; pad estimates and expect surprises.
The suggestions — what they mean and why they matter
Underpriced for time — the slowest jobs bring in about the same money per unit as the fastest, even though they eat far more labor. Why it matters: you’re paying more to do them but charging the same — consider a time-based price, a minimum charge, or a surcharge for the hard ones.
Unpredictable — the time swings widely from job to job. Why it matters: you can’t reliably plan a day around it, so techs run late and jobs get bumped. Pad the estimate and don’t over-schedule this type.
Long tail of slow jobs — most are quick but a handful drag on for a long time. Why it matters: those few outliers wreck the schedule and the average. Find out why — hard access, scope creep, or a tech who needs help — and fix the cause.
The planning numbers
Units/Job — the typical size of one job (median units serviced). A job with 200 extinguishers is far more work than one with 20, so this shows the usual size behind the times.
Units/Day (best · expected · safe) — how many units a route could service in a 420-minute day if work goes fast (best), typical (expected), or slow (safe). We count units, not jobs, because a “job” isn’t a fixed amount of work — units/day stays honest whether a day is big jobs or small ones. Plan staffing with expected, promise customers with safe. Excludes drive time, so real numbers are a bit lower.
Job Median — the typical whole-job time, for putting one job on the calendar. Bigger jobs run longer — that’s expected. To compare how efficient a job type is regardless of size, use the per-unit box chart, not the whole-job time.
Rev/Day — expected jobs per day × the typical revenue per job. A rough ceiling on what one route can bill doing only this work.
Needs attention (top of the page) — pulls the flagged job types to the front so you don’t have to hunt. Every flag is a suggestion with the numbers behind it — a manager makes the actual pricing, staffing, and scheduling call.
Service Time Distribution
The ground-truth measurement under everything else: how long each kind of job actually takes on site (check-in to check-out), from the invoice. This is on-site work time only — it does not include travel. Click any row to see the size breakdown and travel estimate.
What this tells you (plain English)
What it is
For each type of job, it measures the real on-site time — the clock from when a tech checks in at the customer to when they check out. Nothing else: no driving, no lunch, no shop time. Just hands-on-the-job time.
The columns
Median — the typical job (half faster, half slower). The number to trust.
Mean — the average; if it’s well above the median, a few long jobs are stretching it.
Per-unit (in parentheses) — minutes for one unit (e.g. one extinguisher). This is the fair way to compare a 4-unit job to a 100-unit job.
P25–P75 — the middle half of jobs land in this range. A tight range = consistent; a wide range = unpredictable.
Std Dev / Range — how spread out the times are, and the fastest-to-slowest seen.
Diverge — how often the job the office scheduled wasn’t the job that was actually done. High divergence means dispatch is booking the wrong type, which throws off planning.
Open a row to see
Size breakdown — the same job split into Small / Medium / Large by unit count, with minutes per unit for each. This answers “do big jobs get cheaper per unit?” Setup time is fixed, so a 4-unit job is usually far more expensive per unit than a 100-unit job — useful for minimum charges and pricing.
Travel estimate & door-to-done — the typical drive time to this kind of job (measured as the gap from the last job’s check-out to this one’s check-in), plus on-site time, for a rough total per job. This is an estimate — the gap is mostly driving but can include short waits, and travel really depends on the route, not the job type.
What it solves: quoting to the real median instead of guessing, catching scheduling mismatches (divergence), knowing which small jobs are unprofitable per unit, and — with the size and travel views — seeing the true cost of a job door to door.