Checklist for AI lead scoring by company size
Stop wasting outreach on wrong-sized companies. Define specific employee count ranges and organizational signals in your AI lead scoring profile today.
Get startedShort answer
To filter LinkedIn leads by company size, replace vague labels like 'mid-market' with specific employee count ranges in your AI lead profile. Then set hard minimum and maximum scores so the AI stops rewarding a good job title at a company that cannot buy. Test the profile against your best customers and a list of known bad fits before you run outreach.
Define your hard boundaries
Founders often use terms like 'mid-market' when setting up AI scoring. That's a mistake because AI interprets those terms differently by industry. A mid-market software company might have 50 people, but a mid-market manufacturing firm has 500. If you leave the label vague, the score reflects the label, not the company.
Find the exact point where a company is too small to afford you or too large to buy without a six-month procurement process. If you sell payroll automation, a 2-person company doesn't have enough employees to feel the pain of manual entry. If you sell boutique consulting, a 10,000-person company likely has a global agency contract that blocks you.
Find the window where the problem is painful enough to pay for, but the decision maker is still reachable via a direct message. Write that window as numbers. Use employee count first because it is the most visible size signal on LinkedIn. Then add one or two structural signals, such as whether the company has a dedicated department for the problem you solve.
A hard boundary is not a preference. It is a rule the AI applies before it reads the rest of the profile. If a lead fails the boundary, the score should be 0 and the reason should say which boundary failed. That makes bad fits easy to spot when you review a batch.
Vague parameters
- Target mid-sized agencies
- Avoid very small companies
- Focus on established businesses
Concrete parameters
- Score 0 if company size is under 11 employees
- Score 100 if employee count is between 20 and 150
- Score 0 if company size exceeds 500 employees
Translate sizes into AI parameters
AI scoring works best when you link company size to specific roles. A 'Founder' at a 5-person company is a different buyer than a 'Founder' at a 100-person company. In the larger company, they're likely a CEO who delegates buying to a VP. In the smaller company, they may be the only person who can approve the purchase.
Tell the AI to look for markers of growth in your profile. If you target the lower mid market, look for the first dedicated hire in a department. That's a stronger signal of budget than a headcount number alone. A 30-person company with a head of customer success is more likely to buy a customer success tool than a 30-person company where the founder still handles every account.
Use the table below as a starting point, then adjust it to your product. The target role matters because the same headcount can hide very different buying processes. A 60-person agency and a 60-person software company may have the same employee count but different approval paths.
Small (1 to 10)
- Company size
- Founder led
- AI focus marker
- Solo practitioner
Lower mid (11 to 50)
- Company size
- First hire for role
- AI focus marker
- Head of department
Upper mid (51 to 200)
- Company size
- Dedicated budget
- AI focus marker
- Director or VP
Enterprise (201 and up)
- Company size
- Global or regional
- AI focus marker
- Corporate leadership
Build your scoring logic
Your AI instructions should act as filters. Use a hard zero. If a lead fails a primary requirement like company size, the AI shouldn't look at their job title or experience. It should immediately score them as 0 and state the reason. This keeps a strong title from pulling a bad-fit company into your campaign.
Once size is verified, add a complexity check. This checks that even if the company is in your size range, they have the infrastructure to use your product. If you sell HR software, a 40-person company without a dedicated HR person is a bad fit. A 40-person company with an HR manager and an open recruiting role is a better fit.
Write the rules in order. The first rule that matches should win. If the first rule is 'under 11 employees, score 0', a lead at a 6-person company never reaches the rule that rewards a founder title. That order is what stops the AI from averaging a good title against a bad company size.
- 1
Set a minimum employee count
Instruct the AI to assign a score of 0 to any lead where the company has fewer than your minimum threshold, for example 11 employees.
- 2
Set a maximum employee count
Instruct the AI to assign a score of 0 to any lead where the company exceeds your maximum threshold, for example 500 employees, to avoid procurement dead-ends.
- 3
Add a complexity check
Tell the AI to look for signs of a specific department or a job title that only exists in companies that have scaled past a certain point.
Paste this job
Score each lead from 0 to 100 using these rules: 1. If company headcount is under 11, score 0. 2. If company headcount is over 500, score 0. 3. If headcount is 20 to 150 and the person is a founder, head of department or VP, score 90 to 100. 4. If headcount is 11 to 19 or 151 to 500, score 50 to 70 unless the role owns the budget for this problem. 5. If headcount is unknown, score 30 and list 'headcount unknown' as the reason.
Verify with a sample set
Don't trust an AI profile until you test it against known data. Run a small batch of profiles through the system and check the results manually. Start with your best existing customers. If the AI doesn't score your current best clients as 90 or 100, your parameters are too restrictive.
Then find a list of companies you know are too small and ensure they're filtered out. Do the same with companies that are too large. If a 4-person company gets a 70 because the founder has a strong title, your minimum boundary is not working.
If you see too many 'almost fits' (leads just outside your range getting high scores), tighten your employee count boundaries by 10 or 20 people. If your best customers cluster at the low end of your range, move the minimum up. If they cluster at the high end, move the maximum down. Test again before you run a campaign.
Example
- You
I noticed you're scaling the operations team at Acme Corp. Most teams your size struggle with X. Is that happening for you?
- Prospect
We are actually only 4 people, so we just handle it manually for now.
- Suggested reply
Note: This lead was scored as a fit but is too small. Adjust the AI profile to set a hard minimum of 10 employees to avoid this.
When to change the scoring profile
A scoring profile is a snapshot of your current sales motion. It goes stale when your product, price or buyer changes. Review it when one of those things changes, not on a calendar.
Look for these signals that the profile needs a change:
Change one boundary at a time. If you move the minimum and maximum together, you won't know which change fixed the results. Keep a copy of the old profile so you can compare a batch of leads against both versions.
- Your best customers keep scoring below 80, which means the profile is too narrow
- Leads just outside your range keep scoring above 70, which means the boundaries are too loose
- A new product or price point moves the buyer from a founder to a department head
- Your sales team keeps rejecting leads that the AI scored as strong fits
Doing this with Kinlead
Kinlead finds people on LinkedIn who match the customer you describe, or your website, and scores each lead from 0 to 100 against that profile. Leads below the fit line are not kept. Leads are kept in lead groups, and any campaign can point at any group. You can review and adjust the group before outreach starts.
Lead search uses LinkedIn's standard search, so a Sales Navigator seat is not required. You can build a group for a specific size range and use lead groups to keep that range separate from other campaigns.
A scoring profile is only as good as its two hard lines. Test the minimum and the maximum before you trust the score.
Frequently asked questions
Related
Get started with Kinlead
Prefer markdown? Checklist for AI lead scoring by company size.md