The Friday before the Fourth, a loose room. Joe brought a texting platform that any business could use. Melis wanted to sell luxury-hotel rooms direct instead of through brokers. Different problems — but the same idea ran under every answer, applied five different ways.
Stop targeting with the filters somebody else built. Figure out the dimension that actually predicts a buyer, and use Claude to construct the list no vendor sells.
Target the dimension that matters, not the filters that exist
Joe runs an agency and brought a client’s problem: a B2B texting product — a centralized inbox so a multi-location business can text its customers and employees and never drop a thread when someone’s shift ends. Good inbound, no outbound motion, and the classic horizontal trap: any business can use it, which means none of the off-the-shelf filters tell you who to sell.
The fix is to get to the core of the business, because this is a targeting problem. A dentist doesn’t make much more money by texting — nice-to-have, low stakes. But a 24/7 flood-repair company wins or loses the deal at two in the morning; nobody says “we’ll let the house flood until nine.” Their business lives and dies on response speed, so they need texting far more than the dentist does. So does anyone the government compels to keep records — a materials company moving high-dollar cement to the right job site, an insurer required to log communications, a bank. Anytime the law forces a behavior, that’s a good place to sell.
The way to rank it is the delta: the upside of a fast response minus the downside of not having one. For the dentist, the upside is a slightly nicer experience and the downside is nothing. For the flood-repair shop, the upside is closing the deal and the downside is losing it. For the regulated company, the downside is a legal problem. Sort your market by that delta and the answer falls out. “Any business that’s 24/7” is a real dimension — and it’s not a checkbox in any database. You derive it.
"That delta is how you target." — Jordan Crawford
Where you start determines where you end
Here’s the counterexample I gave, because it’s the cleanest version of the principle. I’m building a list of every dentist in the country. I did not start with Google Maps. I started with the NPI registry — every provider is required to have a National Provider Identifier — and then filtered by an active DEA license, because you have to keep that current to administer anything. NPI is free. The DEA match is free. Together they give me the name of every live dentist, which no “dentist” filter anywhere will hand you.
I never would have built it that way if I hadn’t started from the right question: what’s the source closest to the legal reality of this business? Start from Google Maps and you end with a pile of “random businesses.” Start from the licensing root and you end with named people. Where you begin decides where you end.
"Where you start determines where you end." — Jordan Crawford
Begin with your imagination — and invert
The reason people miss all of this is that they’re stuck asking “what tools and filters did we have before?” Wrong frame. These models don’t just do things you’ve thought of; they can discover things you never knew were possible to discover.
So I tell customers: begin with your imagination. Assume you have perfect information on your prospect and perfect information on your customer — you could know anything in the world — and ask what you’d say. Nobody thinks like that, and when you do, Claude comes back with “I can get you eighty percent of the way to perfect.” Eighty percent. From “we have these four filters” to “assume you can know anything” is the whole shift.
Charlie Munger’s version is: always invert. As a young man he managed weather reports for the military, and he framed his job by its worst case — who dies if my forecast is wrong? Three kinds of people. So all he had to do was be certain about the three situations that could kill someone. Same move on Joe’s texting problem: don’t ask “who might like SMS,” ask “where is SMS existential — where does a company get shut down or sued if it doesn’t track its messages?” Invert to the failure, and the target list writes itself.
"Begin with your imagination. Ignore what you think is possible." — Jordan Crawford
Who Gets This
The worked examples below are the actual builds from the room — what we made, what I told each person to do, and the exact moves.
Free: the shift — target the dimension that matters, start from the root, build the filter that doesn’t exist.
$50/mo (most readers start here): everything below the line in this post, and every paid article.
$2,499/yr: Every tool I ship. Edge Copilot is how you talk to all of it through Claude Code. Current tools: Edge Copilot, AutoClaygent, Agent 7, Who to Target and What to Say, Blueprint Cloud, Technology Finder, Video List Extractor, Competitor Monitor, LinkedIn Engagement, Domain & LinkedIn Finder, Dossier Builder, PDF Contact Finder, TAM Contact Harvester, Find a Rep, Blueprint Playbook, Crawford, JoJo. Whatever ships next is included. Plus all 3 courses + weekly Applied Office Hours. (Go annual — $2,499/yr.)
Below the line in this post:
Build your own BuiltWith for fifty bucks — the namespace trick that found 106,000 users
The permit play: mining public records into a $50M upsell list no vendor can sell
Sort the world by any dimension — the luxury-hotel ICP nobody has a filter for
Every question from the room, with the answer
Every week I run Applied Office Hours on Zoom — bring what you're building and we'll work it live.
