Build a Target Account List From Your ICP, Step by Step

To build a target account list from an ICP, translate each criterion into a filter, collect candidate domains from several sources, enrich them with company data, apply disqualifiers, score what remains and sort accounts into tiers. Then add contacts for the buying committee. The profile does the choosing; the list is its output.

Step 1: turn criteria into filters

Take each line of the ICP and rewrite it as something a database or a human can check.

ICP line	Filter
“B2B software”	Industry equals software, not consumer apps
“20-200 employees”	Headcount between 20 and 200
“North America”	Country in US, Canada
“Uses a CRM”	CRM detected or confirmed
“Recent sales hire”	Open or filled sales leadership role in 12 months

Anything that cannot be reduced to a filter belongs in the review step instead. The firmographic and technographic criteria guide explains how to choose checkable fields.

Step 2: source candidate domains

Use several sources so the list does not inherit one database’s bias.

1. Lookalikes of your best customers. Seed accounts from your ICP are the best starting point.
2. Niche directories and marketplaces where your segment lists itself.
3. Job boards, for situation signals such as hiring in a relevant role.
4. Events and communities your buyers attend.
5. Your own site visitors, if you can identify them. BusinessMCP’s page on website visitor identification describes its approach; read its limits before relying on it.
6. Company search tools, with filters matching Step 1.

The free BusinessMCP ICP generator includes a “where to find prospects” section in its output, which can seed this list of sources.

Step 3: enrich and verify

A raw list of domains is not yet a target account list. Add the fields your criteria need: industry, size, location, and any tech signals. BusinessMCP’s companies data API takes a domain and returns firmographics; its bulk endpoint handles up to 100 domains per call, and its page says lookups that find nothing are not charged. Its company search can also return up to 25 companies per call from filters such as industry, country and employee range.

Whatever you use, spot-check at least 20 records by hand. If more than a few are wrong, fix the source before scaling.

Key takeaway: enrichment turns a list of names into a list you can filter. Verification turns it into a list you can trust.

Step 4: apply disqualifiers, then score

Run your disqualifiers first as hard filters. Then score the remainder with the scoring model so accounts can be sorted rather than argued about.

Step 5: tier and assign

Tier	Typical share of list	Treatment
1	Top slice	Personalised outreach, senior rep, account research
2	Middle	Sequenced outreach with segment-specific messaging
3	Bottom	Marketing only: content, ads, events

Set the proportions to match your capacity. A common error is a tier 1 so large that no account gets real attention.

Step 6: add the people

Only now add contacts, using the buying committee in your ICP: champion, economic buyer, end user, blocker. Keep the people data in line with the privacy rules that apply to you. BusinessMCP’s companies API states that it returns no people data, so contact sourcing is a separate step with its own compliance considerations; check current rules for your region.

Step 7: keep it alive

• Refresh quarterly, and whenever the ICP changes.
• Remove accounts that were acquired, shut down or outgrew the band.
• Add accounts triggered by new events. The sales-triggers guide at BusinessMCP lists triggers such as hiring, funding and leadership changes that you can watch for.
• Track which tier produces pipeline, and adjust tier sizes.

Common mistakes

• Buying a huge list instead of building a good small one.
• Skipping enrichment and filtering on guesses.
• Not recording the source of each account, so you cannot learn which source works.
• Treating a target account list as static.
• Starting outreach before messaging is written for the buying committee; see buying committee roles.

Worked example of a list build

Take a fictional ICP: B2B software firms with 20 to 200 employees in North America that use a CRM and recently hired a head of sales. The build might go like this.

1. Export 40 domains of lookalike companies from your best three customers’ networks and directories.
2. Add 150 domains from a niche software directory and 60 from job postings that mention sales leadership roles.
3. De-duplicate to about 220 domains.
4. Enrich with industry, size and location; drop 70 that miss the size band or region.
5. Remove 15 that match hard disqualifiers.
6. Score the remaining 135, tier the top 25 as Tier 1, the next 60 as Tier 2, and the rest as Tier 3.
7. Assign Tier 1 to the senior rep and add two contacts per account.

(The counts are illustrative, not benchmarks.) The sequence matters more than the numbers: collect broadly, filter with evidence, and spend human time only on the top tier.

Measuring whether the list works

After a month, compare the tiers. Did Tier 1 produce more replies and meetings than Tier 2? Did any source produce a disproportionate share of good accounts? If Tier 3 outperforms Tier 1, the scoring model is miscalibrated and should be revisited. Record the results beside the list, then feed what you learn back into the ICP. The list is an experiment, and every experiment should return something to the profile.

Sources and hygiene for the list

Treat every source as biased. Directories overrepresent companies that pay to be listed, job postings overrepresent firms that hire publicly, and event lists favour companies with marketing budgets. Mix sources and record where each account came from, so that after a quarter you can see which source produces the accounts that actually reply and close.

Keep a simple data dictionary next to the list: what each column means, where it came from and when it was last checked. When someone else inherits the list, this one page saves hours of guessing. Mark accounts that have already been contacted, the date, and the outcome, so teams do not email the same company twice with different pitches.

Finally, decide up front what happens to accounts that never respond. A sensible default is a long nurture track with a review after six months: either a trigger event has occurred and the account moves up, or it is dropped to keep the list honest. A list that only grows becomes a graveyard, and a graveyard teaches nothing. Pair that discipline with a quarterly look at the ICP itself, as laid out in what an ideal customer profile is, because a list is only as good as the profile that generated it.