Ask a dispatcher what a good board looks like and you'll usually hear about geography: tight routes, no one crossing town twice, techs home on time. Drive time matters. But a board optimized only for miles can still lose money, by sending your best diagnostician to a filter change while a $9,000 replacement lead waits for whoever's nearby.
Good dispatch balances three goals that often conflict: less driving, more revenue and keeping the windows you promised customers. This post covers how to think about those trade-offs, the constraints that matter more than distance, what to do when the day goes sideways, and what to measure.
Three goals that pull in different directions
Every assignment decision trades one of these against the others.
Least driving. Cluster jobs by area, minimize miles between stops. Lower fuel and wear, more time on site, techs who finish on time. The risk is a board that's efficient but poorly matched: the closest tech isn't always the right one.
Most revenue. Put your strongest closers on the sales opportunities (aging equipment, replacement estimates, high-value repairs) and your efficient techs on quick maintenance. The risk is more driving and techs who resent always getting the long trips.
Keep every promise. Arrive inside every window you gave a customer, even if that means a longer route. The risk is lower utilization: you protect windows by leaving slack in the day.
None of these is right all the time. A useful habit is to choose a lead goal each day based on conditions:
- A heat wave with a full board usually calls for keeping promises, because missed windows turn into cancellations and bad reviews.
- A slow shoulder-season day is a good day to optimize for revenue: send your best salespeople to the replacement leads.
- A day with long distances or high fuel prices might lean toward least driving.
- Most days, you want a balance, with hard limits on what can't move.
Constraints come before optimization
Before you optimize anything, the board has to be feasible. These are the rules that a "nearest tech" approach ignores.
Skills and certifications
Some jobs need a specific skill (ductless systems, commercial rooftop units, tankless water heaters, panel upgrades). Some legally require a certification or license: in the US, techs who handle refrigerant need EPA Section 608 certification, and many electrical and gas jobs require a state or local license. Keep a skills and certification list per tech, with expiry dates, and treat it as a hard rule at dispatch.
Shifts and promised windows
A tech who finishes at 5 can't take a 4-to-6 window in the far corner of your area. A promised window is a commitment. Treat both as constraints rather than preferences, and leave enough slack to absorb a normal amount of overrun.
Parts on the truck
Sending a tech without the likely part creates a return visit. If you track truck stock, check it at dispatch: when the assigned truck doesn't carry what the job probably needs, either send a tech whose truck does or plan a supply-house stop before the job. We cover the cost of return trips in our post on first-time fix rate.
Members and priority customers
If your membership promises priority service, dispatch has to keep that promise on busy days. Give member jobs a priority flag the board respects, not just a note.
Apprentices and helpers
The industry is short on technicians. One industry estimate puts the shortfall at about 110,000, with about 23,000 leaving each year (ACHR News). Growing your own through apprentices is one of the few reliable answers, and dispatch is where that succeeds or fails.
- Pair deliberately. Put an apprentice with a senior tech on the job types they're learning, not just whoever has an empty seat.
- Plan the crew as one unit. A two-person job should be assigned to a pair, with both people's shifts respected, rather than two separate assignments that drift apart.
- Graduate in steps. Let apprentices take simple solo work (filter changes, basic maintenance) close to a mentor, so help is a short drive away.
- Don't count them as a full tech when estimating the day's capacity on complex calls.
When the day goes sideways: intraday re-planning
The plan at 7 a.m. is a guess. By 10, a job has run long, someone's truck won't start, and an emergency call has come in. The question is how quickly and how calmly you adjust.
A simple routine when a job runs long:
- Find what's affected. Which later appointments for that tech will now miss their window?
- Look for a swap. Can another qualified tech, with the right parts, take one of those jobs without breaking their own windows?
- If not, call early. Contact the customer before their window starts, offer a new time, and record it. A heads-up two hours ahead is a different experience from a no-show.
- Protect the high-stakes jobs. If something has to move, move the routine maintenance before the replacement estimate or the member's no-heat call.
The goal is to re-plan only the part of the day that's affected. Rebuilding the whole board every time something slips makes techs feel jerked around.
How TradesWork approaches it
The TradesWork AI dispatcher is a planner, not a chatbot. It builds a proposed board from skills, certifications, promised windows, shifts, parts on trucks, job value and memberships, using one of four goal presets: Balanced, Least driving, Most revenue or Keep every promise. When a job runs long, it re-plans the rest of the day.
The AI part has a narrow job. It turns the dispatcher's notes ("Mrs. Alvarez wants Dan, not before 10") into constraints the planner respects, and it explains each suggestion in a sentence. It never assigns a job on its own; the dispatcher decides.
What to measure
Dispatch quality shows up in several numbers at once. Watching just one invites gaming.
- On-time arrival. Arrivals at or before the end of the promised window, divided by all arrivals. Early counts as on time.
- Revenue per tech-hour. Invoiced revenue divided by paid tech hours. This is where good matching shows up.
- Drive share. Drive time as a share of paid time. Track the trend, not a target to push down at any cost.
- First-time fix rate. A board that saves miles but sends the wrong skills or the wrong truck pays for it in return visits.
- Customer-moved appointments. How many appointments you moved after a promise was made, and how many of those customers you told before the window started.
- Board fill. Booked tech time as a share of available tech time. Low fill on busy days usually means a booking or capacity problem, not a dispatch one.
Review these weekly with the dispatcher, and look at the specific days that went badly. One bad Tuesday, examined closely, usually teaches more than a month of averages.
A dispatch checklist
- Choose a lead goal for the day: balanced, least driving, most revenue or keep every promise.
- Keep skills, certifications and expiry dates per tech, and treat them as hard rules.
- Treat promised windows and shifts as constraints, with slack for normal overrun.
- Check parts on the assigned truck before the job.
- Flag member and priority jobs so the board respects them.
- Plan apprentice and two-person jobs as one crew.
- When a job runs long, re-plan only what's affected and call customers before their window.
- Review on-time arrival, revenue per tech-hour, drive share and first-time fix weekly.