How to write a target profile that actually filters

Ask most companies who their customer is and you get something like this: "B2B SaaS companies, mid market, decision makers."

That describes tens of thousands of companies and hundreds of thousands of people. It is not a filter. It is a category.

A target profile is only useful if it can exclude. Here is how to write one that does.

Start with exclusions, not inclusions

It is easier and more accurate to describe who is wrong for you than who is right.

Write down the companies you would decline. Too small to afford you. Too large to move quickly. Wrong region for your support hours. Wrong regulatory environment. Already committed to a competitor on a long contract.

Every exclusion removes thousands of companies from the list before you have described a single good one.

Define size by something observable

Revenue bands are common in target profiles and mostly useless, because private company revenue is rarely public and usually estimated badly.

Use signals you can actually observe:

• Employee count, which is visible and reasonably current
• Number of people in a specific function, often more predictive than total headcount
• Number of office locations
• Funding stage, where applicable

A company with four people in operations has a different problem to one with forty, regardless of what either earns.

Name the role by what they do, not what they are called

Titles are inconsistent between companies. Responsibilities are not.

This matters more than it sounds. Job title and function changes affect a majority of business contacts within a twelve month period, making title the most volatile field in a contact record. (Landbase) A profile built on exact title strings starts decaying the day you write it.

Instead, describe the person by what they own. The person who owns outbound. The person who signs off on tooling under ten thousand. The person who fixes it when data quality breaks.

Then work backwards to the titles that usually hold that responsibility at your target company size. You will find they differ by size band, which is exactly the detail a title based filter loses.

Add buying signals

This is what separates a target list from a prospect list.

A signal is an observable event suggesting a company is more likely to buy now than last quarter. Common ones:

• Recent funding, which creates budget and urgency
• Hiring for the function you serve, which signals investment in that area
• Leadership change in your buyer's function, since new leaders change vendors
• Expansion into a new market or location
• Public complaints or job posts describing the problem you solve

Signals also expire faster than anything else in a record. Behavioural and intent signals go stale within weeks, while firmographics drift far more slowly. (Derrick) A signal you collected last quarter is not a signal.

A company that matches your profile is a candidate. A company that matches and posted three relevant roles last month is a priority.

Write it as a document, not a mental model

The profile has to exist outside your head, because someone else has to build against it.

A workable profile fits on one page and contains:

1. Exclusions
2. Company characteristics, defined observably
3. The role, defined by responsibility
4. Buying signals, ranked
5. Two or three example companies that fit perfectly
6. Two or three that look like they fit but do not, with the reason

Those last two sections do more work than the rest combined. Examples resolve ambiguity that definitions cannot.

Expect to revise it

Your first profile will be wrong in at least one dimension. That is normal, and it is why the profile is a document rather than a decision.

After the first campaign, look at what actually replied and what actually converted. Those are usually different segments. Adjust toward what converted, not what replied.

The profile should get narrower over time, not broader. Broadening is usually a sign the offer is not landing, and more prospects will not fix that.

---

References

1. Landbase, Data Decay Rate Statistics — landbase.com
2. Derrick, B2B Data Quality Benchmarks 2026 — derrick-app.com

---

We write the target profile with you as the first step of every engagement, then build against it. See Growth Partnership.