What 10,000 carrier calls taught us
We listened to a lot of dispatch calls. Most of what we expected to matter didn't.
Before we wrote much of anything, we sat with the recordings. Thousands of hours of dispatchers and brokers, mostly repetitive, occasionally remarkable.
Here is what actually showed up.
Calls are short and mostly dead air#
The median useful call is around ninety seconds. The median elapsed call is closer to six minutes, and the difference is hold music, transfers, and waiting for someone to pull the load up.
This reframed the problem for us. The bottleneck was never how fast a person could talk. It was how many calls one person could be inside at once — which for a human is exactly one.
Nobody negotiates the way the playbooks say#
The advice everyone gives is to anchor high and concede slowly. What we heard was almost entirely a different move: establish that you can cover it today, then ask. Availability was worth more than posture, consistently.
The calls that ended above the posted rate had this shape:
- Confirm the specifics fast — equipment, appointment, weight
- State the commitment plainly
- Ask once, for a specific number
- Stop talking
Step four was the one dispatchers skipped most often, and it cost them.
The information that decides the call arrives late#
Roughly a third of calls surfaced something in the last twenty seconds that changed the answer — a second stop, a lumper fee, a 2am appointment. Any system that decides whether to take a load from the posting alone is deciding on incomplete information most of the time.
What we were wrong about#
We expected accents, noise and cross-talk to be the hard part. They were a problem, and a solvable one.
The genuinely hard part was memory. A broker who called yesterday expects you to know that. A dispatcher who committed to a lane expects the system to hold that commitment against the next decision. Getting a single call right turned out to be much easier than being the same party across twenty of them.