How to decide who to contact on LinkedIn, scoring leads before you write a word

Short answer: write down what a good customer must have (role, company type, market) and what rules someone out, then score every candidate against that from 0 to 100 using evidence on their profile, not their job-title keywords. Contact only the people above a threshold you can tune, keep timing signals ("why now") separate from fit ("why them"), and check every few weeks whether higher scores really reply more.

Most outreach advice is about what to write. But the biggest lever is who you write to: a perfect message to the wrong person gets ignored, and a plain message to the right person often gets a reply. We wrote about the evidence in lead fit matters more than message copy. This post is the how.

Step 1: separate must-haves from nice-to-haves

Write your ideal customer as two short lists.

• Must-haves: the things without which the person can't buy from you. Usually a role that owns the problem and the budget, a company type, a size range and a market.
• Nice-to-haves: things that make a deal more likely but aren't required: a tool they use, a recent hire, an industry you've sold into before.

Add a third list: "don't target". Competitors, existing customers, students, people in markets you can't serve. These should be hard rules, not soft preferences.

Step 2: score the evidence, not the title

A title keyword is the weakest signal on a profile. "Growth" appears in the titles of founders, interns and fractional advisors alike. Read what the profile actually says: what they're responsible for, how long they've been there, what the company does, where it sells. Three rules that save a lot of wasted messages:

• A matching keyword is not a fit. A "fractional advisor" in your target title doesn't own the budget a full-time head of the function does.
• Seniority and a famous logo never replace a must-have. A VP at a big brand without the core requirement is a maybe, not a yes.
• Location matters when your offer does. If you sell into one market, someone outside it with no sign of expanding there is a maybe at best.

Step 3: use a 0–100 score with plain bands

Numbers force decisions. A simple scale:

Score	Meaning
90–100	Excellent fit on role, company and profile evidence
70–89	Good fit, with some uncertainty or thinner evidence
50–69	Adjacent or uncertain
Below 50	Poor fit, or matches a "don't target" rule

Write one sentence of reasoning next to every score. When a lead replies "not relevant", the reasoning tells you whether the score was wrong or the message was.

Step 4: keep "why them" apart from "why now"

Fit and timing are different questions. A job change, a new funding round or a post about the exact problem you solve says now is a good moment; it doesn't make a poor fit a good one. Score them on separate axes, and use timing to decide the order, not whether to contact at all.

Step 5: pick a threshold, then tune it

Contact only leads above a threshold: 60 is a sensible start on the scale above. Too few leads? Check whether the threshold or the targeting is the problem: if many leads score just under the line, lower the threshold; if few people are found at all, the targeting is too narrow.

Step 6: check scores against replies

Every few weeks, compare acceptance and reply rates by score band. Higher bands should reply more. If they don't, your must-haves are wrong, usually because a "must" is really a "nice to have", or the other way round. How to calculate your reply rate has the formula.

How NudgeLink does this for you

We built NudgeLink around this process, so you don't have to run it in a spreadsheet. You describe your business and ideal customer once; NudgeLink searches LinkedIn, and every candidate gets a 0–100 fit score with a written reason before anyone is contacted.

• The scorer follows the rules above: keywords aren't fit, seniority doesn't replace a must-have, location matters, and your "don't target" exclusions are hard limits.
• Candidates are ranked first, scored quickly, and the most promising are read in full (profile, recent posts) and scored again with that evidence, so credits go where they matter.
• Timing signals are scored separately from fit, and leads are re-checked with their latest posts right before the first message.
• You set the threshold per run (60 by default) and see the whole funnel, including the near misses.
• Accepting and rejecting suggestions teaches the next run.

Then it writes each message for that one person, in your voice, for you to approve. See how lead scoring works, the quality threshold and pricing.