How do you map a market before you build an outbound list?

You map a market before building an outbound list by starting in your own CRM. Read who actually took meetings with you, what made them raise a hand and how they described the pain. Then pull the market that matches, clean it in three or four passes, and give every company a segment and a tier.

I restarted my daily go-to-market standup with this as day one and posted it on LinkedIn, with a video and a diagram of the tools at each step. I redo our own market map over and over, because every time I do, it gets a little better.

Why does market mapping start in your CRM?

Market mapping starts in your CRM because there is the ICP you think you have, and then there is what the data in your CRM says. They are not always the same. The CRM shows who actually took meetings and bought, which makes it a better starting point than a list of filters agreed in a meeting.

This is the part most people skip. They open a list tool, type in the industries and headcounts they have always targeted, and start building. The list looks right because it matches the plan, and nobody checks whether the plan matched the buyers.

For a lot of teams it is worth the time to do this first, before building any list, because it changes what every later step looks for.

What should you look for in the CRM?

Look for three things about the people who took meetings with you: what industry they were in, what made them raise their hand, and how they described the pain in their own words. Those three answers set the market's boundaries, the signals worth watching and the language the messages should use.

• Industry. Count where your meetings actually came from. A sector you never targeted on purpose may be there more often than one you have targeted for years.
• The reason they raised a hand. A new hire, a failed tool, a board asking for pipeline. Whatever it was, it is a signal you can watch for in the rest of the market.
• Their words. The phrases buyers used to describe the problem are better copy than anything you will write from scratch.

Our Chief Strategy Officer, Deisy Lewis, does a version of this with closed-won deals and Claude: export the deals, ask what the best customers have in common, and read what comes back.

How do you pull the first version of the market?

You pull the first version of the market wide, using the industries and sizes the CRM pointed to, and accept that it will contain companies you do not want. I start with Clay's audience feature. These days I mostly run it from Claude Code or Codex rather than logging in to the interface, but the result is the same: a raw table of companies.

Pulling wide on purpose is easier than trying to write perfect filters in one go. A filter that is too tight silently drops good companies you never see. A list that is too loose shows you its problems, and the cleaning passes fix them.

What do the cleaning passes check?

The cleaning passes check, one question at a time, whether each company belongs on the map. I usually need three or four. Each pass removes a kind of bad fit that the first pull let through, and records why, so the same mistake is filtered out next time.

The passes I run:

1. Dead domains. A company whose website no longer works is usually not a company you can sell to.
2. Industries I meant to exclude that still squeaked through the first filter.
3. A sales or marketing leader. If nobody at the company owns the problem we solve, there is nobody to talk to.
4. Headcount, headquarters and revenue against our own limits.

The order matters less than doing them separately. One pass per question makes it obvious which rule removed which company, and that record is what lets the next version be better.

How does each company get a segment and a tier?

Each company gets a segment and a tier after a research agent reviews its website. The segment says what kind of company it is, so it gets the right message. The tier says how good a fit it is, so the best accounts get a person and the rest get a sequence.

Reading the site matters because firmographic fields are blunt. Two companies with the same industry code and headcount can sell completely different things to completely different buyers, and only their own pages say which.

Our ICP map is the written version of this: the passes, the tiers and the reason behind every exclusion. The map then feeds everything downstream. The signal-based list watches only the companies on it, and the plays work them in tier order. Building and keeping that map is part of what we do in GTM engineering.

What should you try first?

Open your CRM and list the last twenty companies that took a meeting with you. For each one, write down its industry, what made it raise a hand, and one phrase the buyer used about the problem.

Then compare that list with the ICP you have written down. If they disagree, believe the CRM. Pull a wide first version of the market that matches it, and clean it one question at a time.