Read and correct an ICP match

Every person Embers finds is judged against your ideal customer profile, and the verdict is the biggest lever on their score.

What the match is based on

The classifier reads the structured criteria on your ICP page, not your original description: job titles, departments, industries, locations, company sizes, buying signals, customers you have, and the titles and industries to exclude. It weighs six dimensions:

Four checks run first and end the evaluation with a non-match:

• An excluded title, in the title, the headline, or a current side role.
• A peer or competitor: someone selling what you sell, a vendor in your category, a journalist at an industry publication, or a coach or course creator in your space. For service businesses this needs an overlap in the service itself; a generic word such as “agency” is not enough.
• Someone who works at your company.
• An excluded industry, read from the profile, or inferred from the company, headline, and role when the field is blank.

The classifier also sees how the person engaged and, once the profile is enriched, their work history, skills, education, and certifications. A role started within about 90 days counts as a strong buying signal, and missing data never counts against anyone. Fixed rules then check the verdict: they can turn a match into a non-match when an exclusion literally applies, never the reverse.

Read the verdict

In the leads list the ICP column shows the fit percentage with four bars for the tier: four for Strong fit (80% and up), three for Good fit (60%), two for Partial fit (40%), and one for Weak fit below. A non-match shows Not ICP and a lead not yet judged reads Scoring. Hover for the tier, the reason, and “Marked by you” on a lead you tagged.

The full lead page shows the verdict with its confidence, from 0 to 1:

• Strong: a match at 0.8 or higher.
• Likely: a match below 0.8.
• Unlikely: a non-match below 0.8.
• No match: a non-match at 0.8 or higher.
• Pending classification: not judged yet.

A confident match lifts the score up to 1.5 times, a confident non-match cuts it hard, and an unclassified lead sits at the neutral middle with a small discount. New leads are classified as scans bring them in, a daily pass picks up anyone left unclassified, and saving an ICP change judges every lead again.

Correct a match

Tagging needs Solo or above, which the free trial includes, and any member of the workspace can do it.

What a tag changes

• At once, for that lead. Your verdict wins: the ICP score becomes 100 or 0, confidence becomes full, and the score is recalculated. A lead you mark a lead joins the default lead views, and it passes the ICP half of the gate every destination applies, so it can reach Slack, HubSpot, or a webhook even when the classifier disagreed.
• On the next classification, for near-identical profiles. Up to ten of your tags, five of each kind and those with reasons first, are shown to the classifier as examples. Each applies only to someone with the same role, seniority, and kind of company, never to a whole job family.
• Never, for your ICP. To rule a whole role or industry in or out, edit the criteria and preview the change. A lead you tick Keep on in that preview is pinned exactly as if you had tagged it.

No later classification or rule reverses a tag. Removing a tag returns the lead to unclassified, with the matching discount, until the next pass judges it again.

Troubleshooting

• A good buyer shows Not ICP. The profile is thin or an exclusion caught them. Tag the lead; if many share the cause, edit the ICP instead.
• People you want are vetoed by title. An excluded title also matches inside current side roles, so a broad word can veto more than you meant. Narrow it on the ICP page.
• You sell services and buyers read as competitors. Peer exclusion is strict for service businesses by design. Tag the lead.

Related