Change Orders on Traffic Control Jobs

Ask a traffic control owner how many change orders they issued last month and the number will be small. Ask their crew leaders how many times last month someone asked them to do something different from what was planned, and the number will not be small at all.

The gap between those two numbers is the most reliably uncollected revenue in this trade.

Why this trade leaks changes specifically

Three structural features conspire.

Changes arrive verbally, in the field, at 6 a.m. The person asking is a prime contractor's superintendent or an agency inspector. The person being asked is a crew leader who is trying to get set up. Saying "I will need that in writing" is technically correct and socially difficult, and the relationship the crew leader is managing is a real one.

The changes are individually small. Another two hundred feet of closure. Six more drums. Stay two extra hours. Come back Thursday. None of them feels like a change order. In aggregate over a month they are substantial.

The work happens before anyone could approve it. By the time the office hears, the extra work is done. A change order raised after the fact is a request for money for work already performed, which is a much weaker position than a request raised before.

Add that traffic control frequently sits under a prime rather than contracting directly, and the incentive to keep things smooth is strong. Crews are not being careless; they are optimising for the relationship in front of them.

The five changes that go uncaptured

Scope extension in the field. The closure gets longer, an extra approach gets covered, an additional intersection gets included. Recognisable because the device count placed exceeds the device count planned.

Duration extension. The job that was quoted for one day runs into two. This one carries a compliance dimension as well as a commercial one: the MUTCD's duration categories escalate device requirements as work lengthens, and work occupying a location more than one daylight period up to 3 days is intermediate-term stationary, for which retroreflective and/or illuminated devices are required because the operation extends into nighttime (Section 6N.01, 11th Edition). So a duration extension frequently requires additional devices, which makes it both more clearly billable and more clearly documentable than a vague "we stayed late."

Standby. The crew is on site and cannot work — waiting on a locate, an inspector, a delivery, a decision. Whether it is billable is a contract question. Whether it is a cost is not.

Rework caused by someone else. Setup built, another party changes their plan, setup rebuilt. Two setups' worth of labor, one setup's worth of billing.

Additional devices left on site. Devices staged for a phase that got postponed sit accruing cost and unavailable to dispatch. On rental terms that is revenue; on a job price it is pure leakage. Either way it depends on the deployment record being accurate — see tracking devices by job site.

Capture beats approval

The instinct is to build an approval workflow. That is the wrong first move, because approval requires the office to be in the loop at 6 a.m. and it will not be.

The right first move is capture: make recording a change take under a minute, on the device the crew leader already has out, at the moment it happens.

What needs recording:

• who asked — a name, not "the prime"
• what changed, in one sentence
• when, timestamped rather than recalled
• the rough magnitude — extra hours, extra devices, extra distance
• a photograph where the change is physical

That is a thirty-second entry, and it converts an unrecoverable verbal exchange into a documented event. Whether it becomes a billed change order is a decision the office makes later, with information.

Capturing on the daily ticket is the natural home, because the ticket is already the contemporaneous record of the job and is already signed. A change noted on a signed ticket has been seen by the customer's representative at the time, which is a materially stronger position than an office note. The wider argument for ticket discipline is in preventing daily ticket disputes.

Set a field authority threshold

The question crews actually need answered is: what am I allowed to say yes to?

Most companies converge on a threshold — some amount of extra hours or extra devices below which the crew leader does the work and documents it, and above which they call the office first. The specific number matters less than that a number exists and everyone knows it.

Two additions worth making explicit:

Anything that changes the traffic control concept is never a field decision. Moving the work space, closing a different lane, switching from a shoulder closure to a flagged one-lane operation — that is a new plan and possibly a new permit, not an adjustment. The boundary between adaptation and redesign is covered in reading a traffic control plan in the field.

Anything the crew leader believes is unsafe is a stop, regardless of who asked. That needs saying out loud and backing up the first time it happens, or the threshold policy is decorative.

Turning captured changes into money

Once changes are being captured, the office has three jobs.

Triage daily, not monthly. A change captured yesterday can be discussed with the customer today, while their own people remember it. The same change raised on an invoice five weeks later is an ambush, and it gets treated like one.

Bill what is billable, promptly. Small changes billed on the next invoice with a clear reference to the ticket and the person who requested them are rarely disputed. The dispute rate rises sharply with elapsed time.

Track the unbillable ones anyway. This is the part companies skip and it is where the strategic value is. A customer whose jobs generate ten unbilled hours a month is costing you real money at your current rate — that is a fact for the next rate conversation, and it is only a fact if it was recorded. It also feeds the standby and non-billable reporting described in job costing for traffic control companies.

The estimating loop

The last use of change data is the one with the longest payoff.

If a particular customer's jobs reliably run 15 percent longer than quoted, that is not a series of unfortunate events — it is an estimating input. If shoulder closures for a given prime always grow by an intersection, price the next one accordingly or add an explicit term.

That loop only closes if changes are recorded against the job and the customer, and if someone looks at the pattern periodically. A quarterly review of changes by customer is a short meeting that reprices several jobs. The estimating framework it feeds is in bidding traffic control jobs, and the collection side is in getting paid faster.

The short version

Traffic control changes are small, verbal and immediate, which means an approval-first process will fail and a capture-first process will not. Give crew leaders a one-minute way to record who asked, what changed and roughly how much, on the ticket, at the time. Set an explicit threshold for what they can say yes to. Triage daily, bill promptly, and record even the unbillable ones — because the pattern across a quarter is worth more than any single change order.

Traffic OS puts change capture on the daily ticket the crew is already signing, so the record exists where the billing and the job history already live. Pricing is flat monthly tiers rather than per user, which matters when the people capturing changes are field crews. Bring a month of jobs that grew to a walkthrough and see what was never billed.