Northeast Freight Time one clock, four readings

Against the Dashboard

The standard response to a dwell problem is a dashboard: average dwell by facility, updating, on a screen, with a target and a colour. For a related perspective on workplace measurement, how employees can tell if they are being monitored.

It is worse than no dashboard, and the reason is specific rather than general grumbling about metrics. For broader industry reporting and context, see DAT Freight & Analytics.

What it displays

The average, which describes almost no actual visit because dwell distributions have a long right tail.

From check-in, because that is where the yard system starts, which excludes the queue outside the gate — the segment most likely to be growing.

Aggregated across modes, so drop-and-hook and live load are blended into a number that moves when the mix changes and nothing operational does.

And updated daily, which invites daily attention to a signal that is mostly noise at any single facility's volume.

Four decisions, none of them wrong individually, producing a number that cannot be acted on and looks authoritative.

What it causes

Optimisation of the displayed number. If check-in defines the start, holding trucks outside the gate improves the metric. Nobody decides to do this; it emerges, because that is what happens when one number is visible and the segments behind it are not.

False comfort. An average within target at a facility whose 95th percentile is six hours. The tail generates every charge and every hours-of-service failure, and the screen is green.

And false alarm. A month where the mix shifted toward live load reads as deterioration and is not.

What to put on a screen instead

If something must be displayed, four figures, per facility, per mode.

Median gate-to-gate, not check-in to departure and not the mean.

95th percentile. The tail is where the cost is.

Share of visits over free time. The only number that predicts invoice volume.

And gate-in to check-in. The invisible segment, and the one that reveals a queue being pushed outside.

Update it monthly. At a single facility's volume, weekly movement is noise, and displaying noise trains people to react to it.

The report that beats any dashboard

A quarterly page per facility. Four numbers, the cause list from the three-word notes, and one paragraph saying what changed and what you intend to do.

It is read by someone, once, with context. A dashboard is glanced at daily by several people without context, and glancing produces impressions rather than decisions.

The quarterly page also survives being wrong. A number in a document can carry a caveat about the mix or the sample. A number on a screen carries a colour.

Where continuous display genuinely belongs

Not an argument that no screen is ever useful.

Live yard state. Trucks on site, doors occupied, next appointment. That is operational information for someone acting on it now, and it is a fact about the yard rather than a judgement about performance.

Appointment adherence today. Also actionable today.

The distinction is the same one that applies everywhere: continuous display suits things somebody is about to act on, and suits performance measurement badly — because performance signals are slow, noisy, and become targets the moment they are visible.

The uncomfortable part

A facility that installs a dwell dashboard has usually done so in response to carrier complaints, in good faith, at real expense.

And it will now measure the wrong quantity, from the wrong start point, at the wrong frequency, with a target attached — and will believe the matter is being managed. That is a worse position than having no data, because the absence of data at least leaves the question open.

Four numbers on a quarterly page costs an afternoon and answers it.

The short version