How to build a signal-based outreach list in an afternoon

Most outreach lists are built the wrong way round. Somebody exports a few thousand companies that match a job title and a headcount band, and the team works through it until morale runs out.

A better list is smaller and built the other way round. Start with the companies, check each one for a reason to talk today, and throw away the ones that do not have one.

It takes an afternoon. Here is the order.

Hour one: write down who you actually sell to

Before any tools, answer three questions in writing.

What has to be true about a company for your offer to work? Not "B2B software", which is not a filter. Something like: sells to other businesses, has a sales team of three or more, and is selling a product with a deal size big enough to justify outbound.

Who feels the pain? Name the role. If two roles could own it, pick the one who has the budget.

What does a bad fit look like? This is the question people skip, and it saves the most time. Too small, too big, wrong sales motion, already using a competitor they are happy with.

Write these as a short list you can check a company against in ten seconds. If you cannot decide in ten seconds, the definition is too vague.

Hour two: pull a rough list of companies

You want forty to sixty names to start, because most will be cut.

Sources that cost nothing:

• LinkedIn search, filtered by industry, headcount and location
• Your own CRM, filtered for closed lost deals older than nine months, which are often the best names on the list
• Attendee and speaker lists from events in your space
• Review sites, looking at who reviewed products next to yours
• Your competitors' public customer logos, for companies that clearly buy this category

Paid databases are faster and worth it later. They are not needed for the first list, and starting without one forces you to look at each company rather than trusting a filter.

Put them in a spreadsheet. One row per company. Nothing else yet.

Hour three: check each company for a signal

This is the part that makes the list worth working. For each company, spend two or three minutes looking for one of these:

• Hiring for sales or marketing roles, on the careers page or LinkedIn Jobs
• Funding announced recently, on Crunchbase or their own press page
• A launch or a new market, on the blog or LinkedIn
• A new marketing or sales leader, from LinkedIn announcements
• A visibility gap against named competitors

Add two columns: which signal, and one line of evidence with the link. The evidence column is not admin. It is what you will write the first line from, and it is what stops you from claiming something that is not true.

Companies with no signal come off the list. Not deleted, moved to a second tab to check again next month.

Fifty companies at three minutes each is about two and a half hours, which is why this is the part people skip. It is also the only part that distinguishes this list from an export.

Hour four: find the person and the address

For each surviving company, find one name. Not four. One.

Use the signal to pick. If the signal is a new head of marketing, that is your person. If it is a demand generation role sitting open, it is whoever that role reports to.

Then verify the address. Guessing a pattern and sending anyway is how domains get burned. Use a verification tool, drop anything that comes back as risky or catch all, and keep your bounce rate low. Deliverability guidance and the rates providers treat as acceptable change over time, so check the current advice from your sending platform rather than an old blog post. [VERIFY: source, current bounce rate guidance from your email provider]

If you cannot find a verified address in two minutes, skip the company. There will be others.

What the finished list looks like

Twenty to thirty rows. That is a week of proper outreach for one person, and every row has a reason behind it.

How to write the first line

Use the evidence column and nothing else. The opener has one job: prove in a sentence that you looked.

Weak, because it could be sent to anyone: I was impressed by what you are building at [company].

Better, because it could only be sent to them: You have had a demand gen role open since July. Who is covering pipeline until they start?

Then stop. A first message should be short enough to read on a phone without scrolling, and it should ask one question rather than proposing a meeting, a demo and a calendar link.

Where this goes wrong

The list grows. Someone decides thirty is too few and adds two hundred unchecked names. The signal column empties out and it becomes an export again.

The evidence column goes blank. As soon as people stop writing down why a company is on the list, the openers go generic.

Nobody rebuilds it. Signals age. A list built in January and worked in March is a list of things that already happened.

Every row gets the same message. If the sequence ignores the signal, the research was wasted. The signal has to show up in the first line, or there was no point collecting it.

Is an afternoon a week worth it?

For most small B2B teams, yes, and the honest reason is that the alternative is not free either. Working a list of five hundred unqualified companies costs more hours than building a list of thirty good ones, and it costs your domain reputation as well.

Run it for four weeks and compare reply rates against whatever you were doing before. That is the only comparison that will convince anyone.

Common questions