- Rewrite your ideal customer profile as filters with thresholds, such as 200 to 1,000 employees and a named tool in the stack, so you can count the market before you email it.
- Write a disqualifier list and a no-contact rule first; they remove wasted sends faster than any change to your copy.
- Score each segment's positive reply rate and meeting yield, and cut the segments that don't convert once they have a few hundred contacts.
“Our ideal customer is a mid-market B2B company that values growth.” Most founders have written a sentence like that, and most sales teams have built a prospect list from it. The trouble is that the sentence excludes nothing. Every company with more than fifty employees fits it, and so does nearly every company that has a website. A profile that cannot exclude anyone cannot guide a list, because a list is nothing but a set of decisions about who to leave out.
A useful ideal customer profile is a set of filters you can apply in a lead tool while watching the count shrink. This guide shows how to turn vague traits into countable filters, how to choose thresholds that mean something, how to add the disqualifiers most teams forget, and how to test the profile against replies rather than against opinion.
Why “mid-market SaaS” is not an ICP
The phrase “mid-market SaaS” describes a category of company, and categories are useful for conversation but not for targeting. Ask three people on your team what it means and you will get three different answers, usually ranging from 100 to 2,000 employees. None of them can be checked against a database field without interpretation, and interpretation is where the list quietly drifts.
The same problem runs through most profile descriptions. “Values efficiency,” “growing fast,” “uses modern tools” and “decision-maker in ops” all describe something real about the buyer, and none of them can be filtered as written. They are conclusions. The filters are the evidence that leads to the conclusion, and the evidence is what you can act on.
The fix is to ask a blunt question about every trait in the profile: what would I see on a company’s website, in its job postings, or in a database that would tell me this company has the trait? If you can’t answer that, the trait belongs in a sales conversation, not in the targeting document.
| Fuzzy description | Filterable version |
|---|---|
| Mid-market | 200 to 1,000 employees, checked against a current headcount source |
| Growing company | Hired three or more sales roles in the last 90 days, or announced a funding round |
| Uses modern tools | Lists a named tool in job postings or on a public stack page |
| Operates at scale | Has multiple offices, or serves customers in three or more countries |
| Decision-maker in ops | Title matches a short list, such as Head of Sales Ops or RevOps Director |
Firmographics: the filters that narrow the market
Firmographics describe the company as a business: its size, sector, location, revenue band and funding stage. They are the cheapest filters to apply and the easiest to check, which makes them the place to start.
Thresholds should reflect the things that actually constrain your buyers. Company size matters because it tends to predict whether there is a process to adopt a tool, a budget line to pay for it, and a team big enough to use it. A product that takes a dedicated admin to run will fail at a company of twelve people, and a company of 5,000 may have a procurement process that takes nine months. The right band is the one where your product works without a special effort from the customer.
Sector matters for a different reason. Some industries have regulatory constraints that change what your product can do for them, and some have sales cycles that make cold outreach a poor fit. Be specific. “Financial services” is too broad to be useful. “Regional banks and credit unions under 500 staff” is a segment you can work.
Geography is a filter teams underuse. Time zones, languages, and the legal rules that govern cold email all depend on where the prospect sits. A profile that says “English-speaking markets” should be a list of countries, and each country’s rules should be checked before the first send. The regulatory side of that deserves a proper review, because the rules differ by country.
The mechanism is multiplicative. Each filter you add cuts the pool by some fraction, and the fractions compound. The illustration below shows how a profile with four filters reduces a broad market to a list you can work carefully.
Technographics and triggers: the filters that create timing
Firmographics tell you who could buy. Technographics and triggers tell you who is likely to buy now, which is the more valuable question.
Technographics are the tools a company uses. They matter when your product integrates with, replaces, or depends on something else. If your product sits on top of a CRM, a company without that CRM is not a prospect, and a company using a competitor’s tool is either a strong fit or a hard switch, depending on your offer. Public signals are often enough: job postings name tools, careers pages describe stacks, and integration directories show who has connected what.
Triggers are events that change a company’s need. A cluster of sales hires often means a team is scaling outbound and will need tools to support it. A funding round means budget and a reason to invest in growth. A new executive in a relevant function can mean a new mandate. Triggers are perishable, so the window matters. A hiring spike from six months ago is history, and a trigger from the last 90 days is a reason to write this week.
Keep technographic and trigger filters separate from firmographics in your profile, and document what each one predicts. A trigger is a hypothesis about timing, and it should be tested like any other hypothesis. If you can’t say why a trigger would make someone more likely to reply, you are filtering on coincidence.
Buyers and users: who replies, who decides
Teams often write one persona and try to reach everyone in the buying group with it. That doesn’t work well in cold email, because the people who reply are usually not the people who sign. The economic buyer may be a VP who never answers a cold message. The champion, often a director or manager who lives with the problem daily, is much more likely to reply, and to refer you upward if your message gives them a reason.
Separate two levels in your profile. The account-level ICP describes which companies you target. The persona-level description says which roles inside those companies you contact, and in what order. Reach the champion first, ask for the right person if they aren’t the owner, and keep the economic buyer as a later step.
Persona titles also need to be checked against reality. Titles vary by company, and a “Head of Operations” at a 200-person company may own the entire revenue process, while the same title at a 2,000-person company may own a single warehouse. Use title lists as a starting point, then confirm the reporting line on the company’s own pages or public profiles before you assume scope.
Build the profile from closed deals, then check it
The most reliable starting point for an ideal customer profile is not a workshop. It is the record of the customers who already paid. Pull the last 15 to 25 closed-won accounts and write down the observable facts about each: employee band, sector, country, the tools they run, the role of the champion who signed off, and what changed in the company in the six months before they bought. Then look for the patterns that repeat. A cluster of three or four shared traits is worth building into a filter. A single trait that appears once is an anecdote.
Do the same for the late-stage losses, the deals that consumed weeks of time and then stalled. Their traits often reveal a segment that looks like a fit in your filters but behaves differently in practice, usually because the buying process is longer, the budget sits elsewhere, or the problem is one they can live with. A profile built only from wins overstates how easy the market is.
Once the traits are written down, compare them with your filters. Where the filters and the closed deals agree, the profile has support. Where they disagree, one of them is wrong, and the closed deals are usually the better evidence. Be careful with small numbers. Fifteen accounts is enough to see a strong pattern and too few to rule out a weak one, so treat the output as a set of hypotheses to test with campaigns rather than settled fact. Refresh the comparison every two quarters, because the customers you won last year may not describe the market you are selling into now.
Disqualifiers: the list of who you refuse to contact
Most ICP documents describe who to target. The more valuable section describes who to exclude, and it is almost always missing. Disqualifiers do more to protect a campaign than any copy change, because they remove sends that were never going to convert.
A useful disqualifier list has five kinds of entry. First, companies below the size where your product works, however enthusiastic they seem. Second, sectors or regulatory environments you can’t serve, including any where your product’s terms would be a problem. Third, existing customers and active trials, which belong to customer success, not outbound. Fourth, competitors, partners and their staff, which waste sends and invite awkward replies. Fifth, anyone who has asked not to be contacted, or who appears on your suppression list. That last category is a legal requirement in most markets, not a preference.
Add observable exclusions too. A company that has just announced layoffs is usually a poor target for a new purchase, and a company that just reorganized a function may have no owner for the problem yet. Check the news before you sequence. These exclusions cost minutes to write and days to recover from if you skip them.
Write the disqualifiers first: they are easier for a team to agree on than the target profile, and each one removes a category of wasted send before anyone has spent time on copy.
Turning your ideal customer profile into a scoring model
Once the filters are defined, you can score a list to decide who goes into which sequence. A scoring model does not need to be sophisticated. A simple points system covers most of the value.
Give each company points for firmographic fit, technographic fit and trigger strength, and give the contact points for persona match. A workable structure is 0 to 3 points for firmographic fit, 0 to 2 for tool fit, 0 to 3 for trigger strength, and 0 to 2 for persona match. Set bands. A total of six or more goes into the primary sequence, four or five into a lighter sequence or a later wave, and anything below four is excluded.
The scores are guesses until they are calibrated. After a campaign, compare reply and meeting outcomes across the score bands. If the high band doesn’t outperform the low band, the model is not measuring what you think it is, and the weights need to change. If the bands do differ, you have a model worth keeping, and the difference is the most useful thing the campaign taught you.
Keep the model simple enough that someone can explain each score in one sentence. Complicated models are hard to audit, and they hide the assumptions that need testing.
Testing the profile, and letting it change
The test of an ICP is whether its segments produce different outcomes. Run a campaign to a few hundred contacts in each of two or three segments, not to the whole pool, and compare positive reply rate and meeting yield. Remember that small samples produce noisy differences, as the A/B testing guide explains, so look for large, consistent gaps rather than small ones.
Cut the segments that consistently produce “not now” and “wrong person” replies, and expand the ones that produce interested replies. Revisit the disqualifiers each quarter, because the market moves and so do your product and your evidence.
Don't widen the profile after a weak campaign: a poor result usually points to the copy or the offer. A bigger, looser pool dilutes the signal you need to learn from, and it adds sends to the segments that were already failing.
Finally, write the profile down where the team will see it, and link each filter to the evidence behind it. A profile that says “companies with 200 to 1,000 employees” is a rule. A profile that says the same thing and notes that the segment produced a positive reply rate three times the average last quarter is a decision the team can revisit, challenge and improve. That second kind of document is the one that keeps getting better.
If you are building the list now, the guide to building a clean, targeted lead list covers the sourcing step, and a comparison of cold email with LinkedIn outreach helps decide which channel suits each segment. OutboundLeadPro’s lead finder lets you apply filters like these and see the count before you commit, and you can run the campaign from your browser at OutboundLeadPro.
Keep reading
Cold Email A/B Testing: Why Most Tests Prove Nothing
Cold email A/B testing is often too small to mean anything. Here is the sample-size math at real reply rates, and how to read the results.
Daniel Khoury
September 19, 20263 min readHow to avoid spam folder placement: 12 causes and fixes
Want to avoid spam folder placement? These 12 causes, from authentication to list quality to complaints, each come with a concrete fix you can check.
Daniel Khoury
September 24, 20263 min readHow to build a lead list that stays clean after the first send
Build a lead list around a real ICP, with stage gates that reject bad rows before they cause bounces and complaints. Includes a worked example.
Daniel Khoury
October 7, 20263 min readSPF, DKIM and DMARC explained for cold email
The three email authentication records decide whether you reach the inbox — and most cold outreach fails because one is missing or wrong. Here's each one in plain English.
The OutboundLeadPro Team
October 11, 20262 min read