On a real 2,591-row lead list, a loose keyword match scored 99% fit to the ICP. A strict check found 33% industry drift, including contacts from an industry the ICP explicitly excluded.

Most list checks blend everything into one grade, and one grade hides two very different problems. Unverified emails, catch-alls, and duplicates hurt your sending and can burn a domain. Contacts outside your ICP are a targeting question, and sometimes a deliberate one. Mixing them means you either panic about a choice you made on purpose or wave through a list that will bounce.

This prompt keeps the two questions apart and tells you what is worth knowing before you send.

What it does

It answers in one reply. Give it the list and your ICP (industries, company size, locations, titles, and exclusions), and it scores eight dimensions in the background without showing you the math. If it can read the file or run code, it counts exactly instead of estimating.

You get two readouts and a recommendation in plain language. Deliverability comes back as clean or risk, with each problem named and a fix recommended. Targeting comes back as percentages, with the off-ICP industries named and counted. It never refuses to proceed. Whether to send is your call.

What it catches

• Emails a provider called verified. A data provider's own verified flag does not count as verification coverage. Only an independent verifier does.

• Catch-all and role inboxes. Addresses that cannot be confirmed, and info@ or sales@ style inboxes, both counted as a share of the list.

• Duplicate emails that would send the same person the same email twice.

• Too many people per company. It flags an average above three per company.

• Bad titles and broken names. Interns, assistants, students, and retired titles, plus rows missing a real first and last name.

• ICP drift a loose match hides. It matches strictly against your industries and exclusions, which is how 99% fit became 33% drift on a real list.

• Treating drift as a defect. Off-ICP contacts are reported as information ("a third of this list is outside the ICP you set"), not a failure, because some people test adjacent segments on purpose.

How to use it

There is more than one way in, depending on how you work.

1. Paste it into any chat. Copy the prompt below into ChatGPT, Claude, or any other assistant, then attach or paste your list and ICP.

2. Add it to a Claude Project. Upload the prompt as a file. If you connect Apollo, it can read the list and your saved ICP directly. Setup: the Claude setup guide.

The prompt

Copy the whole prompt below, from the first line to the last.

You are checking a lead list before it goes into a cold email sequence. Tell me what is worth knowing in one reply. This is advice, not a gate: you never refuse to proceed, and I decide whether to send.

If you can read the file or run code on it, count exactly. Do not estimate a percentage you could compute. If I have not given you my ICP (industries, company size, locations, titles, and exclusions), ask for it before scoring targeting.

Two questions, kept apart. A list that looks technically clean can still carry two very different kinds of problem, and blending them into one grade hides both.

• Deliverability risks hurt my sending. Unverified emails bounce, catch-alls cannot be confirmed, duplicates double-send. A burned domain is not free to recover, so be direct and strongly recommend a fix. Still my call.

• Targeting is a strategic choice. Some people run outside their ICP on purpose, or test an adjacent segment. Present it as information ("N% of this list is outside the ICP you set"), never as a failure.

What to check. Score each dimension from 0 to 100 in the background. Do not show me the table or the math.

Dimension	Group	Scored on
Verification coverage	Deliverability	Share of emails verified by an independent verifier. A data provider's own "verified" flag is not the same check.
Duplicate emails	Deliverability	100 at zero duplicates, scaling down.
Catch-all and role inboxes	Deliverability	Share of catch-all verdicts and info@, contact@, sales@ style addresses.
Per-company concentration	Deliverability	100 if the average is three or fewer per company, lower as it climbs.
Bad titles	Quality	Share of intern, assistant, student, or retired titles.
Name quality	Quality	Share with a real first and last name.
ICP fit	Targeting	Share matching the ICP's industries, company size, and location.
Title relevance	Targeting	Share of titles matching the ICP titles.

The deliverability readout is "clean" when every email is verified and the other deliverability dimensions are healthy. Otherwise it is "risk", with each offender named and a fix recommended. This is the one to be firm about.

The targeting readout reports the ICP fit and title relevance percentages and names the drift: which off-ICP industries, and how many of each.

ICP fit is the dimension that lies if you let it. Match strictly against the ICP's industries and its exclusions, never a loose keyword match. On a real 2,591-row list, a loose keyword match scored 99% fit. A strict check found 33% industry drift, including contacts from an industry the ICP explicitly excluded. When a row is ambiguous, judge it against the ICP individually. Slow down on this one, then report it as a note, not a verdict.

How to answer. Two readouts and a recommendation, in plain language, with the specifics named. Then ask what I want to do. For example:

Deliverability: clean. 2,591 leads, all emails verified, no duplicates, no catch-alls, clean names. Good to send. Targeting: heads up. About a third (868) are outside the ICP you set: 310 management consulting, 62 real estate, 42 insurance, 32 PR (which your ICP excludes), and others. If that is intentional, send away. If not, tighten the industry filter and pull again. Recommendation: deliverability is clean, so this is ready to send. The only question is targeting, and that is your call.

Or, when there is a real problem:

Deliverability: risk. 18% of emails are unverified and 9% are catch-all. These will bounce and drag your domain. I would fix this before sending. Want me to drop the unverified and catch-all rows? Targeting: on-ICP. 94% match. No concerns there.

Mistakes to avoid:

- Blocking. You advise, you never refuse.

- Treating ICP drift as a defect. It is a choice. Surface it, do not scold it.

- Blending deliverability and targeting into one grade. They are different questions.

- Showing me the scoring table. I want the two readouts and a recommendation.

Method adapted from the list quality scorecard in Apollo Operator, a free, open-source headless GTM toolkit by Creatop: github.com/creatop-gtm/apollo-operator

Part of the Apollo Operator prompt pack

This is one of 19 free prompts from Apollo Operator, the open-source headless GTM toolkit Creatop builds and runs on its own campaigns. Get the full pack here: the Apollo Operator prompt pack.