Campaigns gather all the data. Then they don't use it.

LOUISVILLE — Every campaign I've ever walked into has a folder. A voter file, a turnout model, a precinct targeting list, maybe a poll. Somebody paid real money for all of it. And in race after race, it sits there — opened once, admired, and then quietly ignored while the campaign goes and does what it was always going to do anyway.

That's the part that costs you. Not the data you didn't have. The data you had and didn't use.

A targeting list you don't cut turf from is just an expensive apology you wrote to yourself in advance.

Here's what it looks like in practice. The model says these forty precincts are where your persuadable voters actually live. The campaign mails the whole list anyway, because that's what the printer quoted and nobody wanted the fight. The early-vote report says your base is banking ballots a week ahead of schedule, so you could move that last $30,000 from a TV flight into a knock program that turns out the soft supporters the model already named — and instead it goes to one more week of an ad that's reaching people who voted on day one.

Or take the canvass — the one that stings the most, because the data is the richest and the waste is the worst. Your volunteers knock doors for months. At every single door they're collecting gold: who's a supporter, who's undecided, and what that voter said was the one issue keeping them up at night — street by street, neighborhood by neighborhood. By October you can see that the doors in one part of town keep naming the same issue while the doors three miles away name something else entirely. That's a message map and a turnout map, built for free by your own people. In most campaigns it dies in the app it was typed into. Nobody pulls it. Nobody lets it change the mail, the script, or the where. The knockers head back out the next morning to collect more data no one will ever read.

And here's the part I'll only say halfway. Most campaigns tag those doors on the old 1-to-5 support scale — strong yes down to strong no — and that scale tells you which way a door leans and almost nothing about what to do with it. There's a sharper way to ID a door: one that tells you the action, not just the lean. I don't give that one away on the internet. Reach out — or, if you've been one of my clients, you already have it.

None of that is a data problem. It's a decision problem. The number was sitting right there.

I've spent over twenty years watching good operatives lose to this, and it almost always shows up in the same two places: the mail budget and the last ten days. Mail, because "send it everywhere" feels safe and "send it to the forty precincts that matter" feels like a risk you'll have to defend. The last ten days, because that's when fear takes the wheel and the plan goes out the window.

The fix isn't more data. Most campaigns are already drowning in it. The fix is deciding, before the money's committed, that the number gets a vote in the room. That when the model and the gut disagree, the model at least gets to make its case. That every line in the budget can answer one question: what in the file told you to spend this here?

That's the whole job, really. Not gathering the numbers. Using them — especially when using them means doing less, in fewer places, than the version of the campaign that just wanted to feel busy.

I read the numbers straight because the alternative is expensive. Ask anyone who found the right precinct list in a drawer the week after they lost.