LinkedIn Sales Navigator Search Filters That Narrow to Your ICP

A Sales Navigator search returns a result count, and the count feels like progress. The decision underneath it is narrower: which filters define who belongs in the list, and which ones only decide what you look at first. Confuse the two and you lose the ability to tell a targeting problem from a copy problem, because the list is a different list every week.

What Sales Navigator Actually Gives You to Filter On

LinkedIn's lead and account filter definitions organise lead filters into seven groups: Company, Role, Personal, Buyer intent, Best path in, Recent updates and Workflow. Account filters sit in three groups: Company attributes, Spotlights and Workflow.

The split between those two filter sets matters more than the individual names, because several attributes that sales teams write into an ICP document exist on only one side of it.

If revenue or headcount growth appears anywhere in your qualification rules, a lead search cannot enforce them. The working order is to build the account search first, save the result as an account list, then run a lead search filtered by that account list using the Workflow group. Teams that skip the account step usually end up approximating revenue with headcount, which quietly folds a professional services firm of 40 people and a logistics firm of 40 people into the same segment.

Build Membership on Attributes That Survive the Quarter

Our rule is that a filter belongs in the membership test only if it will still be true in 60 days. Company, Role and most Personal filters pass. Buyer intent, Following your company, Viewed your profile recently and Posted on LinkedIn in the last 30 days do not.

Behaviour filters are genuinely useful, but as a priority order rather than a boundary. Put them in the membership test and the population changes underneath you between one sending week and the next, so a drop in reply rate has at least two candidate explanations and no way to separate them. Keep membership stable and the same drop has one explanation worth investigating.

The exception is an ICP genuinely defined by an event. If you sell transition services and only companies that hired a new operations leader in the last quarter can buy, then Changed jobs becomes the membership rule and the firmographic filters become the sort. That inversion is deliberate and worth stating out loud, because the list then has to be rebuilt on a schedule rather than maintained.

Function and Seniority Beat a Long Title String

Job titles are not standardised across companies, so a Boolean title list silently drops every buyer whose employer uses a different word for the same job. LinkedIn's Function and Seniority level filters are its own normalisation of that mess, which makes them the more reliable primary role filter.

The trade-off is real: Function is coarse and will pull in adjacent roles you did not want. You pay for reliability with a longer list, and the sampling step below is what keeps that cost visible. Use title Boolean for exclusions, where precision helps and a miss is cheap. Avoid it for inclusion, where a miss is invisible.

The Boolean Ceiling Shapes How You Phrase a Query

LinkedIn documents Boolean search on Sales Navigator with three operators, AND, OR and NOT, which must be typed in uppercase. Precedence runs quotes, then parentheses, then NOT, then AND, then OR. Wildcard searches with an asterisk are not supported, and Boolean works in the Company, title and keyword fields only, not across every filter. LinkedIn's Boolean query limitations page states a maximum of about 15 operators in a single query.

Two consequences follow. Without wildcards you cannot write one term that catches a word and its variants, so every variant costs you a separate OR and eats into the ceiling. And 15 operators is not enough to enumerate a real title landscape, which is the practical argument for pushing inclusion into Function and Seniority.

A query that respects both looks closer to this than to a 40-term list:

That is five operators, leaving room for the company-side terms in the same search.

Prove the List Before It Enters a Sequence

Before a list reaches a sequence we sample 25 profiles from the result set, open each one, and answer a single question: would I be comfortable if this person received the first email tomorrow? Fewer than 20 yes answers means the filter is wrong, and no amount of rewriting fixes it.

When the sample fails, the failing profiles usually tell you which filter to change:

1. The company fits but the job does not, which means the Function or Seniority setting is too broad or a title exclusion is missing.
2. The job fits but the company does not, which usually means the account-side filters are doing too little and the search should have started from an account list.
3. Both fit on paper but the market is wrong, because Geography is set by region when the buyer's territory is narrower, or because the industry label on LinkedIn does not match what the company sells.

Record the sample result with the campaign. A list that passed at 22 of 25 and then underperformed is a copy or timing question. A list that was never sampled is not evidence of anything.

Saved Searches Are a Monitoring Layer

LinkedIn's guidance on saving lead and account searches states a cap of 50 lead searches and 50 account searches, and says you will receive weekly emails and homepage alerts when new people or companies match the criteria.

Both numbers carry a decision. The cap means a saved search slot has to be earned, so retire the ones nobody has acted on rather than collecting them. The weekly cadence means a saved search is a monitoring tool: if your motion depends on reaching someone within a day of a job change, the alert arrives too late and the trigger has to come from somewhere else in the stack.

Start With the Sample, Not the Filter Panel

Open your current Sales Navigator search, pull 25 profiles, and count the yes answers. That number tells you whether to touch the filters at all, and it takes about twenty minutes. A pass means your next experiment belongs in the copy or the timing. A fail means every copy test you run this month will measure the wrong thing.

If you want the targeting rules, the list build and the sequences running as one system your team owns, book your free ICP and campaign-fit discovery call →.