How to Define an Ideal Customer Profile for B2B Outbound
Write your ICP as filterable rules, build it from closed-won deals and add a timing signal to decide who to contact first. See the steps, scoring example and tiers.

Write your ideal customer profile as rules you can filter a database by, such as industry, employee band, region and tools, then add a timing signal to decide who gets contacted first. Base the rules on your best closed-won customers, not on a paragraph of adjectives.
What is an ICP for outbound?
Rework describes three layers that change at different speeds:
- Account fit: size, industry, geography and tools, which rarely change.
- Contact fit: the department, seniority and title you write to.
- Timing: signals that an account might buy now, such as a hire or a funding round.
Commercify says a profile written as a paragraph, such as "innovative mid-market tech companies," cannot build a list. Its checkable example: B2B software companies, 50 to 500 employees, using HubSpot, with at least two open sales roles.
How do you build one?
Using Shipyard's method and Commercify's advice:
- Pull your closed-won deals from the last 12 to 18 months and list shared traits. Favor customers that closed fastest, paid most and stayed longest.
- Study closed-lost deals too. Who did not buy is often as useful as who did.
- Pick 4 to 6 fields that separate winners. Drop any field where won and lost accounts look the same.
- Weight them to add up to 100.
- Test on 50 accounts and compare the ranking with how your best rep would rank them by hand.
- Recalculate every quarter with new won and lost deals.
With no customers yet, start with a narrow hypothesis and treat your first hundred conversations as a test of it.
| Dimension | Example weight | What it captures |
|---|---|---|
| Firmographic fit | 40 | Industry, headcount, revenue, region |
| Buying triggers | 25 | Funding, leadership change, hiring surge |
| Technographic fit | 20 | A tool that predicts need |
| Engagement | 15 | Site visits, replies |
Source: Shipyard's example weights. It says to adjust them to your own closed-won data.
How do you tier accounts?
Rework's tiers:
- Tier A: fit plus a live signal. Personal, researched messages across several channels.
- Tier B: fit, no signal. Short sequences, with signals rechecked monthly.
- Tier C: near-fit accounts, used only in small test batches.
Rework's rule: timing never overrides fit, so a hiring spike at a bad-fit company is still a bad fit. Shipyard's example cutoffs are 80 to 100 points for A, 55 to 79 for B and 30 to 54 for C, adjusted until each tier's close rate really differs.
Which signals matter?
Apollo layers them: firmographic fit as the gate, technographic fit to qualify, intent to prioritize, and trigger events such as funding rounds, new executives, job postings and product launches to accelerate. It also applies suppression rules before any account reaches a sequence.
Who inside the account should you email?
Belkins' 7.5 million emails show founders replying at 0.57% against 0.42% for C-level, and companies of 0 to 10 employees at 0.72% against 0.22% at 10,000 or more. Higher reply rates do not mean better customers, so weigh them against deal size.
Rework warns that technographic filters are hypotheses: "no CRM detected" means none was visible on public pages. Spot-check 20 accounts by hand before pulling 2,000.
Next, build and verify the list with how to build a B2B email list, add personalization, and track results against reply rate benchmarks. For who runs outreach, see SDR vs BDR vs appointment setter.
The Sales Room builds targeted lists from your ideal customer profile. Book a free strategy call.
Frequently Asked Questions
- What is an ideal customer profile for outbound?
- A set of filterable rules for account fit and contact fit, plus timing signals that show who may buy now.
- How do I build an ICP without customers?
- Start with a narrow hypothesis and use your first hundred conversations as a test of it, per Commercify.
- How often should I update my ICP?
- Quarterly. Shipyard recommends recalculating weights with new closed-won and closed-lost deals.