The rundown is a PDF the panel can actually use

Mamba Smith is the founder of Queen City Garage, which produces The Grand National Show: a weekly NASCAR panel show on Grand National Today where broadcasters and drivers argue about the last race and pick the winner of the next one.

The rundown for The Grand National Show is one PDF, sent to every person on the panel the day before we record. It has the show's logo on it, the recording link at the top, last week's picks up front, and four topics per show. It is written for a producer and a guest at the same time.

It took two rejected drafts to get there, and the rejections were the useful part.

The two that did not ship

The first drafts this week were plain text files. They had the right topics. They also had no branding, buried the picks halfway down, left out the recording link, and read like notes to the editor rather than a page for a guest.

The note back was direct: where is the branding, these need the recording link, the picks go up front, and write it for the people receiving it. A rundown is not an internal document that happens to be shared. It is the first thing a guest sees about the show that week. If it looks like a text file, the show looks like a text file.

What the PDF is now

One branded page per show. The logo. The recording link and the call time, in the right time zone, spelled out. Last week's picks with the result beside each one, because that scoreboard is the cold open and every panelist wants to know where they stand before the light goes on. Then four topics, each with one question and the clips we hold to support it, so a guest knows what is coming and a producer knows what to cue.

For this week's two shows the page went through six versions before it went out: the four-topic structure, guests marked as sitting in for a regular, two seats corrected, the Tuesday 11 a.m. Eastern call time, and two rounds of tightening. Five participant emails went out from the show's own address once the last version was approved.

A producer's version, with clip timecodes and a say-this-not-that list, exists separately and goes to nobody on the panel.

Why it is built from the record

The rundown is generated from the episode record rather than written fresh. The episode number, the topics, the picks and the panel come from the same rows the website reads. That is the simplest way to guarantee the PDF a guest is holding on Tuesday says the same thing the episode page says on Friday.

Before recording, our software also reads the race transcripts and coverage we already hold, matches them against the topics, and flags candidate clips with timecodes. This week it surfaced 49. A person chose which ones the rundown would promise.

The lesson, which is not about PDFs

The text-file version was faster to make and correct in every fact. It was still wrong, because it was built for the person making it rather than the person reading it. Most of the automation in this series exists to give a person time. This is the post about what the person is supposed to do with it: look at the thing from the other side of the table.

What a person did: rejected two drafts, set the structure, wrote the questions, chose the clips. What the machine did: pulled the picks, the panel and the topics from the record, flagged 49 clip candidates, rendered the page.

This is No. 09 in How We Make the Show, the Queen City Garage build log on producing The Grand National Show. Watch the show at Grand National Today.