📦Trade press - carriers publish on-time data for the first time

Several carriers have begun publishing a single monthly on-time percentage each, voluntarily and in public. A competitor has said it will pull those figures into its product.

Why it matters for us. The question of the discovery is whether customers can score their carriers fairly on data Halden already holds. This article shows the unfair version arriving first, and free.

What is wrong with one self-reported number.

• Who counted. Each carrier defines on time for itself. Within the hour or within the day is the difference between 97 and 81.
• What was carried. A carrier running short urban lanes and one running cross-border freight are not doing the same job. Comparing their percentages compares their lanes. That is 🧩 Confounding, and Hale gave it a whole Tuesday.
• How many. A small carrier's month is a few hundred loads. Its number will jump about from luck alone, and a ranking will read each jump as a change in quality. See 🧩 Regression to the mean.

What I did with it. Took it to Thursday's one-to-one and to Zainab Karimi as a gift: a public example of the thing our prototype has to be better than. Our advantage is exactly what the published page lacks. We see every load, with its lane and its timestamps, from the shipper's side.

Three weeks later a carrier-relations manager at one of our design partners made the same points from the other direction, with more force and less statistics: 📞 Gideon's objection. He thinks his small carriers will be the ones punished, and he is probably right about the public page. Whether he is right about ours is the work of November.

Position on Stowline. They will be first, with borrowed numbers. I would sooner be second with ranges.

Where it went, by late November. From 3 November every number I report at work carries its spread, as 🧪 Report every metric with a range for a month. The weighting analysis turned the complaint into a proposal, 🔍 Show carrier scores as ranges, not ranks. And the part that is not statistics at all, what a fair score would owe a small carrier, is being thought about slowly in 🏔️ On fairness in rankings.

• A fictional trade-press item, from a weekly that does not exist. It landed on Wednesday 1 October, three weeks into 📂 Carrier Scorecards discovery, and three people forwarded it to me before nine.
• A group of mid-size road carriers has started publishing its own monthly on-time delivery figures, on a shared page, for anyone to read. The piece says it is the first time carriers in this market have done so without being asked by a customer.
• Each carrier reports one percentage per month. No lane, no volume, no definition of on time that I could find in the article or behind it.
• The trade paper treats it as a transparency story. A consultant is quoted saying shippers will finally be able to compare.
• 🏢 Stowline gets a paragraph: its carrier-management product will apparently import the published figures directly.
• My first reaction was that someone had shipped our project. My second, about an hour later, was that they had shipped the version we are trying not to build.