How to find prospects on LinkedIn: search by persona, not by job title

Ask a room of salespeople how to find prospects on LinkedIn and most will describe the same move: type a job title into the search bar, add a country, start sending requests. It is fast, and it produces the wrong list. It misses the buyers who describe their job in other words, and it fills up with people who share the title but will never buy.

The fix is to start from the buyer rather than the title: write down who you are looking for as a sentence, turn that sentence into a short search, and look at every profile before it becomes a lead.

Why a job-title search is the wrong way to find prospects on LinkedIn

A LinkedIn people search matches the words people put on their own profile. That makes a single title a poor proxy for a buyer, for three reasons.

Titles do not line up across companies. The person who owns lead generation at a 40-person software company might be “Head of Growth”, “Marketing Manager”, “Demand Generation Lead”, or a co-founder who never bothered with a title. Search one of them and you miss the other three.

Titles attract people who share the word, not the job. Search “Sales Director” and you also get recruiters hiring sales directors, consultants who coach them, assistants to them, and people whose headline still says it two jobs later.

A title says nothing about the context that makes someone a buyer. Company size, business model, the problem they are accountable for this year: none of it is in the title, and all of it decides whether your message lands.

Write the persona as a sentence first

Before you touch the search bar, write one sentence that a colleague could use to pick the right profiles out of a pile. It needs four parts:

1. Function and seniority: who owns the problem, and who signs.
2. Company context: type, size, market.
3. The situation that makes them need you now.
4. Who it is not: the look-alikes you want out.

Compare the two:

• Weak: “Marketing Director”
• Strong: “Marketing leads at B2B software companies of 20 to 200 people who own a pipeline target, not agencies, not freelancers, not recruiters.”

The last clause does more work than it looks. Most of the noise in a prospect list comes from one or two look-alike groups, and naming them up front is how you keep them out.

Write the sentence in the language your buyers use on their profiles. A sales head in Lyon writes “directeur commercial”, not “sales director”. If your market is bilingual, you need one search per language.

Turn the persona into a short boolean search

LinkedIn’s people search understands boolean operators: AND, OR and NOT in capitals, quotation marks for exact phrases, and parentheses to group alternatives. A persona sentence translates into three building blocks:

• One group of title variants, joined with OR.
• One context group, added with AND, when a skill or market word really separates buyers from look-alikes.
• One exclusion, with NOT, for the biggest source of noise.

("Head of Marketing" OR "Marketing Director") AND ("SaaS" OR "B2B") NOT "agency"

Keep it short. A string with a dozen nested groups tends to come back empty or strangely random on a standard account. Two or three groups of two terms each, plus one exclusion, is a sensible ceiling, and several short searches beat one monster query: one per family of titles, one per language.

Precision also saves you from a LinkedIn limit. A free account that searches people heavily, the way prospecting or hiring does, eventually hits the “commercial use limit”, which caps search until the first of the next month. Ten precise searches cost far less of that allowance than fifty vague ones.

Review every profile before it becomes a lead

A search result is a candidate, not a prospect. Look at the headline, the company and the location of each profile, and decide in a few seconds whether it matches the sentence you wrote.

Write the reason down in one line, for the ones you keep and the ones you drop:

• “Runs marketing at a 60-person software company, owns the pipeline number.”
• “Agency founder who sells to marketers: out.”

Those reasons pay off twice. The reason you kept someone is usually the opening of your first message. And the reasons you dropped people tell you whether the persona was wrong: if half the rejections share the same reason, it belongs in the NOT of your next search.

Import only what you can work. Invitations are the real bottleneck: LinkedIn caps them per week (the figure is not published), so a list of a hundred strong fits is weeks of outreach already. A list of a thousand maybes goes stale before you reach the bottom of it.

Note the connection degree as you go. First-degree contacts can get a message straight away; second- and third-degree need a connection request first.

Doing this in Swimlane

Swimlane runs the whole routine above from one sentence, and still leaves the decision about who gets in to you.

Describe the persona and run the search

1. Fill in Settings → Personalization → Your ideal customer and How you help first. Every profile the search finds is rated against those fields, so an empty field means a meaningless rating.
2. On your dashboard, under Find new leads, open People Search.
3. Describe the persona in a sentence, exclusions included, and set the maximum number of leads (15 by default, up to 100).

The AI turns your sentence into the kind of query described above: job-title variants grouped together, a skill group when it matters, and at most one exclusion, in the language you wrote in. A single word is searched as it is. The search runs on LinkedIn from your own connected account.

If it comes back empty, broaden the persona or split it into two searches rather than retrying the same sentence.

Review the fit, then import to a list

The review screen lists every profile with its photo, headline and location, best matches first. Each one carries a colour-coded match against the ideal customer you described, and hovering it shows the one-sentence reason.

You then choose what to import: the profiles you ticked (Import selection), everything (Import all), or every profile matched above 50%. Ticking by hand is worth the minute it takes, because the reasons are where you spot the look-alikes.

Swimlane creates a new list named after your persona and opens it. The contacts arrive with their headline, company and location, plus their connection degree whenever the search result showed it. That degree matters later: a campaign only sends a connection request to a contact whose degree it knows, and a contact of unknown degree leaves the flow at that step. Imports count toward a ceiling of 2,500 contacts a day, and the plan holds 5,000 contacts per seat.

Feed a campaign from the list

1. Create an Outbound Campaign and pick the list in the Audience step.
2. Turn on ICP match. Swimlane rates each contact on five levels, from very low to excellent, against the ideal customer written for that audience (pre-filled from your settings). Set Match strength to High match or better to keep only strong fits.
3. Start the flow with an If connected condition. Yes goes to a message; no goes to a Connection Request with its Wait for acceptance window. A first-degree contact that reaches a connection request step leaves the flow, so this condition is what keeps your existing connections in the campaign.

Enrolment is continuous: contacts added to the list later join the running campaign on their own. Each new People Search creates its own list, so move the good fits into the campaign’s list with the Move to list bulk action. Invitations go out within your daily ceiling (20 by default, adjustable up to 200), spaced at random intervals.

Mistakes that ruin a persona search

• Searching one title and calling it done. Run the variants, and one search per language.
• Writing the persona without exclusions. The look-alikes you do not name end up in your list.
• Importing everything the search returns. Volume you cannot work is not pipeline.
• Trusting a match rating against an empty profile. Write your ideal customer first.
• Sending the same message to every result. You wrote down why each person fits; use it.

Where to go next

A persona search finds people by who they are. To find them by what they are talking about, pair it with finding prospects by the topics they post about. Once the list exists, the connection request templates help the first touch get accepted, and the evergreen campaign on a living list keeps a list like this one working without relaunching anything. For the full routine around it, see the SDR prospecting workflow.

The other sourcing channels are on the LinkedIn prospecting page, and the five fit levels are explained on the ICP fit page.