# How to choose who to target with cold email

Companies don't buy; people in companies do. Strong outbound starts from groups of people who share a situation, not from five famous logos or one ICP written at the start. Here is how we map them, and how you can do it too.

Canonical: https://slsbmb.com/guides/who-to-target/ · Updated 2026-10-30 · SLSBMB AI

## Companies don't buy, people do

You can email one person at a company and get a shrug, email another and get a $10,000 contract, and email a third at the same company and get $100,000, for the same offer. There is no abstract "Company X". There is someone at Company X who wants a bonus for modernizing the infrastructure.

So we research audiences: groups of people in a shared situation, with one reason to reply, one clear exchange with you and one path to a decision. The industry is only one of their traits. The CTO of a marketplace who has to push every fix through the app store is an audience. So is the head of lead handling at a car dealership whose leads go cold between the CRM and the call.

One technology can live in several audiences. A delivery marketplace and a restaurant with its own app need a similar solution, but their processes, budgets and decision-makers differ, so their emails must differ too. The reverse also holds: if only the job title changed while the problem and the offer stayed the same, you haven't found a new audience, just another way to find people in a known one.

## The real reason to buy often lies elsewhere

In our agency years, a maker of sensors for truck fleets believed parking was its main selling point. Customers actually bought because of cargo theft, since the sensor alerts when a driver breaks procedure, and because of safety rules at loading. Fundraising software sold badly to charities even with a strong number behind it: one large charity had grown its donations three and a half times with it. Recipients recognized a salesperson at once. After studying dozens of replies we saw that these people were proud of their role, not of the money, and one opening line about pride in helping organizations that save lives changed the replies. Finding it took hours of discussion. Audiences like these can't be found by industry name, only by understanding people.

## Map the market in two directions

- **Inside out:** from your real customers, capabilities and history to the people they are useful to.
- **Outside in:** from the way the market divides itself to the parts where you have something to offer.
- **Intermediaries:** integrators, agencies and platforms that have their own reason to introduce you to their clients. Each is an audience with its own benefit, and the email to an intermediary is written differently from the one to a buyer. "He knows a lot of people" is a weak reason to write; a paid task they close, a client they help or a wider offer they gain is a strong one.

The best intermediaries can be unexpected. An accounting service entering the UK hit a wall: everyone there already has an accountant around the corner. The clue came from small YouTube channels about accounting with a few hundred views: if people watch them, they are unhappy with something. Nobody offered those creators anything, and almost all of them answered a partnership offer. Judge intermediaries by money, not by signed agreements: for one client a single deal with an end buyer once brought ten times more than fifteen reseller contracts, because resellers sign gladly and resell almost nothing.

## A spectrum, not a list

We see a market as a continuous spectrum. In the center are the obvious buyers; further out, less obvious but plausible uses; further still, reasonable experiments; past the last border is fantasy, where we don't go. Never lose the center: a researcher working for hotel software who didn't find the hotels has failed. But don't stop there either: markets the client never asked about have often performed as well as the core ones in our campaigns.

Every serious idea gets a fate: its own audience, part of another, a reasoned rejection or an open question for research. That is why the number of segments says little by itself. Fourteen well-researched audiences can cover more of a market than thirty industry names, and cutting thirty down to fourteen mechanically is an easy way to throw away half of it.

## Turn audiences into searches a database can run

A market speaks with many names. One type of company calls itself a marketplace, a booking platform, an ordering app or an on-demand service; one responsibility is called Head of Mobile, VP Engineering, Director of App Development or CTO depending on the industry and country. Search one name and you find a small slice of the market, then conclude the market is small. We call the cure coverage conscience: if the best search lane in a big market turns out tiny, someone's name is missing, so keep looking.

Our playbooks search along four stable axes: what the company does, the person's job title in English, the country of the headquarters and the approximate company size. Each audience unfolds into a matrix of company descriptions × job titles in every spelling: 31,847 combinations for the 14 audiences of the [published Revopush playbook](https://slsbmb.com/demos/market-scans/?case=revopush#explore), 13,151 for LeadToCall's 30 and 8,718 for the 11 audiences of our own fundraising playbook.

- **Skip technology and revenue filters.** In contact databases those fields are often empty, outdated or guessed, and a filter on them quietly throws out good buyers whose field is blank. Let the email explain who the offer is for.
- **Write to the top as well.** In small companies that is the CEO, COO and founders; as companies grow, add marketing, distribution, sales, business development and product. Executives know what their company exists for, and a smart email gets forwarded to whoever owns the result.
- **Don't confuse proxies with facts.** A company's headquarters country is not every country it operates in, and its headcount is not its number of app users.

## Size the market honestly

Estimate each audience from the bottom up, through company types, roles and the overlaps between them, and count through parent markets so one person found under three titles isn't counted three times. The estimate is a map of the territory, not a list of contacts already bought: how many usable contacts a campaign gets becomes clear when contacts are purchased.

A top-down figure such as "the AI sales market is worth $X billion" is useless for outreach, because you can't write a single email from it. Filters in a contact tool that shrink with every condition are misleading the other way. Across the playbooks we have delivered, the average market map holds 1.8–5.6 million people; even a Sending Engine at $10,000 a month would need 3.5–10.6 years to write three emails to each of them. For most businesses the market runs out later than the budget.

## How the playbook does this for you

The [AI Outreach Playbook](https://slsbmb.com/pricing/) is this whole method, run by specialist AI agents over hours of work and 30–47 million tokens: the audiences with an article each (who they are, why they need it, the evidence and sources, priority and size), the emails, and the search matrix. A chatbot answers "who should I sell to" in a few thousand tokens, and you get a list of names; see [what that difference looks like on one brief](https://slsbmb.com/compare/chatbot/). Research at this level from a large consulting firm costs hundreds of thousands of dollars ([RocketBlocks](https://rocketblocks.me/guide/business-model.php)); the playbook costs $500 at the launch price.

## Frequently asked questions

### How do I figure out my ideal customer profile for outbound?

Start from situations, not firmographics: groups of people who share a problem, a reason to reply and a path to a decision. Map them from your real customers outward and from the market's own divisions inward, add intermediaries, give each one evidence and a size, and then translate each into company types and job titles a database can search.

### How many audiences should I target at once?

As many as the evidence supports in the research, and as many as your budget can test properly in sending. On a starter budget, two or three strong audiences are enough; switch the others on as the engine grows, and they stay in your playbook meanwhile.

### Should I filter my list by technology or revenue?

Usually not. Those fields are often empty or outdated in contact databases, so the filter drops good buyers silently. Search by what the company does and the person's role, and let the email say who the offer fits.

### Should I email the CEO or the manager who owns the problem?

Both. The executive knows why the company exists and forwards a smart email to whoever owns the result, which is one of the shortest ways in, while the manager feels the problem every day.

## More guides

- [How to choose an outbound provider](https://slsbmb.com/guides/choosing-an-outbound-provider/)
- [What cold email costs in 2026](https://slsbmb.com/guides/cold-email-cost/)
- [Cold email deliverability without risking your domain](https://slsbmb.com/guides/cold-email-deliverability/)
- [How to calculate cold email ROI](https://slsbmb.com/guides/cold-email-roi/)
- [How a cold email sequence works](https://slsbmb.com/guides/cold-email-sequence/)
- [Cold email vs LinkedIn, ads, events and calls](https://slsbmb.com/guides/cold-email-vs-linkedin/)
- [Email warmup: why 21 days, and what real warmup looks like](https://slsbmb.com/guides/email-warmup/)
- [The first 90 days, step by step](https://slsbmb.com/guides/first-90-days/)
- [How to get your first B2B customers with outbound](https://slsbmb.com/guides/first-b2b-customers/)
- [Is cold email legal?](https://slsbmb.com/guides/is-cold-email-legal/)
- [What to do when a prospect replies](https://slsbmb.com/guides/replying-to-prospects/)

---

Start with a free interview: https://slsbmb.com/#start · Pricing: https://slsbmb.com/pricing/ · Contact: a@slsbmb.com