Building a Sales Navigator lead list that doesn't waste your send limit

Why a lead list is a spending plan, not a database

A prospect list is not an asset you accumulate. It is a queue of decisions about how to spend a limited number of connection invites, and LinkedIn meters those per account per week. Once an invite goes to the wrong person, that slot is gone until the window resets. Nothing about the list being bigger gets it back.

Do the arithmetic on your own numbers and the point becomes uncomfortable. If your sending account can realistically place a fixed number of invites in a week, and 15 of every 100 rows on the list are people who should never have been on it, then roughly one working day in seven is spent inviting the wrong people. Over a quarter that is weeks of capacity. The fix is not more volume, it is a cleaner list.

This reframing changes how you treat the build. Most teams optimise for list size because Sales Navigator reports a result count and a big number feels like progress. The number that matters is the share of rows a reasonable person would defend if you opened them one by one. If you are also deciding how much sending capacity you need in total, our notes on LinkedIn lead generation walk through the capacity side; this post is about what you put into it.

The list does not cost you money. It costs you invites, and invites are the scarce thing.

Stack your filters in this order

Build the list from the company inward. Account-level filters first, person-level filters second, behaviour filters last. Doing it in the other order makes every later filter work on a population you have not yet agreed you want.

1. Start in account search, not lead search. Fix headcount band, industry, and the geography where the company actually operates. Save that as an account list.
2. Add the account-level signals that matter to your offer: recent funding, hiring in the function you sell to, headcount growth in a specific department. Signals are where account search earns its price.
3. Only now move to lead search and scope it to your saved account list.
4. Filter people by function and seniority rather than typing job-title strings. Titles are written by whoever was hiring that month; function and seniority survive the variation.
5. Add tenure filters. Someone three weeks into a role has no budget yet, and someone eleven years in may have stopped answering strangers.
6. Apply activity and recency filters last, as a trim rather than a foundation.
7. Save the result as a lead list, not a saved search, so the rows freeze while you clean them.

The function-and-seniority point is the one that changes results most. A search for the literal phrase "head of demand generation" will miss the person carrying that job under four other titles, and will catch a contractor who put it in their headline. Function plus seniority plus company size is a description of a job, and that is what you are actually targeting.

Resist the temptation to add every available filter. Each one narrows the list, but each one also introduces a way for LinkedIn's data to be wrong about someone. Three well-chosen filters with a manual pass beat eleven filters with none.

The exclusions most lists forget

Filters decide who gets in. Exclusions decide who quietly should not have, and they are where most of the waste hides because nothing in the interface prompts you to think about them.

The contacted-log is the one worth building properly. It needs nothing more than profile URL, date, which profile sent, and outcome. Without it, two campaigns run six weeks apart will overlap, and the overlap is invisible until someone replies asking why you keep writing.

The QA pass: twenty-five rows by hand

Before the list leaves your hands, open a random sample of rows and judge them as a human would. Twenty-five is enough to tell a clean list from a dirty one, and small enough that nobody argues about doing it.

1. Pull a genuinely random sample, not the first page. The first page is sorted by relevance and flatters the list.
2. Open each profile and ask five questions: is this a real person, is this their current employer, does this job actually own the problem we solve, are they in the region we can serve, and would a colleague be a better fit at the same company.
3. Mark each row pass or fail with a one-word reason.
4. Count the failures. That count out of twenty-five is your junk rate for the whole list.
5. Fix the filter that produced the failures, rebuild, and sample again. Do not hand-delete the bad rows and ship the rest.

The last step is the one people skip. If four of your twenty-five failed because the employer on the profile is out of date, deleting those four leaves the same proportion of stale rows across the other several hundred you did not open. The sample is a measurement of the filter, and the filter is what you repair.

Write the junk rate down each time you build. It becomes the fastest signal you have that a segment is decaying, and it lets you argue for a rebuild with something better than a feeling.

What the person sending actually needs from you

A clean list handed over badly still produces bad sending. The sender needs enough context to write a note that sounds like it was meant for one person, and nothing beyond that.

• Profile URL, full name, current company, current title
• The segment tag that explains why this person is on the list, in plain words rather than a code
• One relevance field per row: the signal that got them included, such as a recent role change or a specific job posting at the company
• Which sending profile owns this row, decided before sending rather than during
• A do-not-contact column, populated from the CRM and from anyone who has previously asked to be left alone

Assigning rows to sending profiles in advance matters more than it looks. A profile whose network is a coherent slice of one industry keeps looking like a real professional in that industry, while a profile whose connections span six unrelated segments stops resembling anyone. If sending runs across managed or rented profiles, that segmentation is part of how the arrangement is kept sane, which we cover under LinkedIn account management, and it also changes how many profiles you need in the first place — see how many accounts to run for outreach.

How often to rebuild

Rebuild on decay, not on the calendar. People change jobs, companies restructure, and a list that was accurate in January will quietly stop being accurate without telling you. The junk rate from your QA sample is the trigger: when it drifts meaningfully above where it sat on the first build, the list has aged out.

In practice that means keeping a short standing routine rather than a big quarterly project. Re-run the account search, diff it against the saved account list, and look at what left and what joined. Additions are usually the most valuable rows you will get all month, because a company that just crossed into your headcount band or just started hiring for the function you sell to has a reason to talk that did not exist before.

One more discipline: keep your segments small enough that you can describe each one in a sentence. If a segment needs a paragraph to explain, it is two segments, and the note you write for it will be vague for both halves. Our broader approach to B2B lead generation on LinkedIn starts from the same premise.

Key takeaways

• Every row on the list consumes one invite from a weekly allowance you cannot carry over, so list quality is a budget decision.
• Filter from the company inward: account filters, then function and seniority, then activity as a trim.
• Exclude current customers, existing connections of the sending profile, dormant accounts, region duplicates and anyone inside your own cool-off window.
• QA twenty-five random rows by hand, count the failures, and fix the filter rather than deleting the bad rows.
• Assign rows to a specific sending profile before sending so each profile keeps a coherent network.

Frequently asked questions

Do I need Sales Navigator to build a usable list?

No, but you will work harder. Regular LinkedIn search hits a commercial-use limit on heavy searching and lacks account-level signals like headcount growth and hiring activity. If you are building lists weekly, the account-search-first workflow described here is difficult to reproduce without it. Check LinkedIn's own pricing page for current costs rather than trusting a figure in a blog post.

How big should a lead list be?

Size it to your sending capacity for the period, not to the search result count. Work out how many invites your accounts can realistically place per week, multiply by the number of weeks the campaign runs, and add a small margin for the rows QA will remove. A list several times larger than that will go stale before you reach the bottom of it.

Should I clean the list or let the sender skip bad rows?

Clean it first. A sender skipping rows in the moment is making judgement calls with no record, which means the same bad rows reappear in the next build and nobody learns which filter produced them. Cleaning centrally is also faster, because one filter change removes hundreds of rows at once.

What do I do about people who changed jobs since the list was built?

Treat a job change as a new row rather than an update. The relevance signal that justified including them has changed, the company no longer matches your account filters, and the note you planned to send no longer makes sense. Move them to a separate list, decide whether the new company qualifies, and write for the new situation.