How to prioritize LinkedIn prospects by ICP fit, not a points score

Every team that tries to prioritize sales prospects on LinkedIn ends up building a points grid. A profile visit is worth this much, a click that much, a reply more, and everything decays after a few weeks. Three months later nobody opens the dashboard, because the “hottest” leads are a recruiter, a competitor and a student who clicked on everything.

The weights are not the problem. The question the grid tries to answer is. This guide replaces it with two simpler questions — does this person fit, and is something happening in the conversation — and shows how to run your day on them.

Why points-based lead scores break on LinkedIn

Lead scoring was designed for marketing automation, where a download or a visit to the pricing page says something about intent. LinkedIn activity says much less.

• Cheap signals dominate. A like, a profile view or a click costs the other person nothing. The people who produce the most of them are often the least likely to buy: job seekers, vendors pitching you, peers from your own industry.
• Weights need maintenance that nobody does. Every new offer, market or campaign changes what a signal means. The grid stays frozen on the day someone built it.
• A number invites false precision. Is a 62 better than a 58? The team ends up debating the model instead of writing messages.
• It blends two questions. “Should I sell to this person at all?” and “Is this the moment?” need different answers. Adding them into one total hides both.

Fit first, timing second

Split the two questions and each one becomes easy to handle.

Fit is about who the person is: their role, their seniority, the kind of company they work for and the market it sells to. It is stable. It changes when they change jobs, not when they like a post.

Timing lives in the conversation: they accepted your invitation, they replied, they looked at your profile, a thread has been waiting on you for three days. Those are not points to add up. They are events to act on, in the inbox, today.

The working rule: fit decides who deserves your effort, the thread decides when. A reply from an average fit gets an answer this morning. An excellent fit who has never heard of you gets a considered first message, not a bump on a leaderboard.

Write your ideal customer as a paragraph, not a grid

A plain-language paragraph beats a table of criteria because it can say what matters most and what rules someone out. Four things make it work:

1. Name the person, not only the company. “Heads of sales and sales ops leads”, not “SaaS companies”.
2. Give the context. Industry, company type and size, how they sell: “B2B software companies of 20 to 200 people whose team prospects on LinkedIn.”
3. Write the restrictions out loud. “Not agencies.” “Outside financial services.” “Non-technical founders.” Restrictions do most of the work: someone who matches the role but breaks a restriction is usually a peer or a competitor, not a buyer.
4. Say who is out. Students, recruiters, freelancers, consultants — whoever keeps showing up in your lists and never buys.

Compare two versions:

Vague: “Decision-makers in tech companies who care about sales.”

Useful: “Heads of sales, VPs of sales and sales ops leads at B2B software companies of 20 to 200 people whose team prospects on LinkedIn. Not agencies, not recruiters, not freelance consultants.”

The second one can be checked against a profile. The first one matches half of LinkedIn.

Five fit levels instead of a number

Once the paragraph exists, sort people into a handful of levels rather than a score: very low, low, medium, high, excellent. Five is enough to decide what to do and too few to argue about. Tie each level to an action:

• Excellent and high: a message you write yourself, about something real in their situation. There are few of them; they deserve the time.
• Medium: a sequence with a good template, and a closer look as soon as they reply.
• Low and very low: out of outbound. Keep them in the CRM, because people change jobs, but do not spend invitations on them. LinkedIn caps how many invitations you can send each week, so every one that goes to a poor fit is one a good fit does not get.

Review the mediums once a week. That is where your paragraph is wrong: a role you forgot to name, a restriction you never wrote down.

How to prioritize sales prospects day to day

A routine that holds up when the inbox is full:

1. Conversations waiting on you come first, whatever the fit. A live thread outranks any cold prospect. If the backlog is the real problem, start with a routine to get your LinkedIn inbox to zero.
2. Then the good fits you have not engaged yet. Pick a few, read their profile, write one real first message each.
3. Let campaigns handle the middle. Set the campaign audience to medium fit or better and let the sequence send invitations and first touches at a human pace.
4. Close the loop. When a deal is won or dropped, move the contact to a closed stage so they stop coming back into your shortlist.

Doing this in Swimlane

Swimlane is built around this split. There is no points grid to configure and no score to read.

• Where you write it. Settings > Personalization > “Your ideal customer” holds your default. In a campaign’s Audience step, switch on ICP match: the field starts from that default, you can rewrite it for this campaign, and Match strength sets the floor — Low, Medium or High match or better.
• How a contact gets a level. Swimlane breaks the paragraph down into the roles that define your buyer, the context around them, the terms that contradict a restriction and the hard exclusions, with variants in English and in your own language so a profile written in French still matches. Each contact in the audience then lands on one of five levels, from very low to excellent, based on their headline, current role, company, About section and your private notes. Market words alone (“SaaS”, “B2B”) cannot lift someone above medium: the role has to be there. An exclusion drops them to very low.
• It stays current. Edit the paragraph, or let a profile change, and the levels concerned are recalculated.
• Where you see it. On the contact panel, with one row per campaign; on contact cards and list rows; and on pipeline cards, which show the best level and each campaign’s level on hover.
• Your daily shortlist. The dashboard block People to interact with lists up to five high or excellent fits you have barely interacted with, leaving out archived contacts and anyone in a pipeline column marked Done or Abandoned. A selector narrows it to one campaign’s profile.
• When you source. A People Search written as a persona shows each profile on the review screen with a one-sentence reason it fits or doesn’t, judged against what you wrote in Settings > Personalization.

The ICP scoring page shows the whole feature, and a flow that gives your best fits a shorter, more personal sequence is one way to act on the levels.

Where fit stops being enough

A fit level only knows what is written on a profile. A founder with a one-word headline and an empty About section will land low even if they are exactly your buyer. Two habits fix most of it:

• Write what you learn in a private note. “Signs off on sales tools budget” is text the level takes into account.
• Trust the conversation over the level. A medium fit who replies with a real problem is worth more this week than an excellent fit who never answers. The level orders your effort; it does not close deals.

Fit tells you where to look. What happens next is still a conversation.