A recruitment agency can have hundreds of LinkedIn profiles and still have no useful employer-client audience. The list may mix internal talent leaders with agency recruiters, candidates commenting on vacancies, protected clients and managers whose hiring needs have already closed. Another export of the same broad search does not resolve those differences.
Build the audience around the desk you sell, the employers you can serve and the people whose current responsibilities relate to that hiring problem. This guide gives you an original employer-client brief and a replayable CSV exercise. The work ends with a separate accepted audience, an account count and a remaining delivery requirement. Candidate sourcing after winning an assignment belongs in the recruiter sourcing-brief guide.
Define the employer-client desk before searching
Start with a service you can explain concretely: the role families you place, the employer types you understand and the territory you cover. For an industrial desk, “Northeast US engineering employers hiring project managers or engineers” is more useful than “companies that are hiring.” Record what makes your service relevant, using experience you can substantiate.
In a January 2026 firsthand agency discussion, a life-science recruiter described slow client decisions and relying on LinkedIn messages. In a March 2026 business-development thread, a solo operator reported poor responses despite already having contact-data tools. These accounts suggest a practical question: does the audience express a credible reason for that employer to consider this particular desk? They do not establish a conversion rate or prove that more exported profiles will solve a sales problem.
Use the blank employer-client brief CSV. Its twelve requirements include an illustrative desk rule and an empty actual_desk_agreement column. Complete the real agreement, name the reviewer and version it before collection. This is a planning sheet, separate from your operational prospect CSV.
| Boundary | What to specify | What changes the decision |
|---|---|---|
| Employer fit | Sector, relevant hiring function, territory and account exclusions | A staffing supplier or a protected existing client is outside this new-client job. |
| Person fit | Current HR, talent, department or founder responsibility | A useful title still needs current remit and employer evidence. |
| Commercial route | Hiring context and the actual route for discussing agency support | An unknown supplier route remains a question; a vacancy does not establish an agency mandate. |
Find employer-side people with current LinkedIn search
Keep the employer shortlist and the person shortlist connected. Review the company and its current hiring context, then search for the relevant people at that employer. For a project-manager desk, that could mean an operations or delivery leader and the internal HR contact. For engineering hires, inspect the engineering leader and the talent function. The brief decides which roles belong; collecting every executive adds rows without proving relevance.
- Record the employer. Save the company reference, territory evidence and the reason its hiring work fits your desk. Distinguish the work location from the headquarters and from a person's location.
- Record a dated hiring observation. Save the relevant role or announcement reference and the date you inspected it. Check whether the role is still open. A funding announcement can motivate research, but it does not prove a vacancy or willingness to use an agency.
- Search People. Type the role and employer context into LinkedIn, choose People and inspect the available filters. Current company and location can help narrow the audience.
- Inspect current responsibility. Verify that the person is internal to the employer and relevant to the department or hiring process. “Talent partner” could describe an employee, an external consultant or a recruitment supplier.
- Save the source with the profile. Keep the public person-profile identity, source reference and observation date so another consultant can understand why the row was included.
LinkedIn's current People-search documentation describes natural-language queries as well as exact Boolean search. In the conversational search experience, a starting phrase could be HR leaders at manufacturing employers in New York. Do not surround a natural-language query with quotation marks. Review the results and refine the company and location filters rather than treating the phrase as an exact job-title rule. LinkedIn also labels search-result counts approximate and its “Actively hiring” filter a Premium feature.
Standard LinkedIn provides a starting route. If you already use Sales Navigator, its documented function, seniority and account-list filters can support a more specific employer search. Keep that seat when those filters serve your desk; it is not necessary to buy one merely to try Leadsforlinked's standard LinkedIn profile extraction. Follow the search-to-CSV workflow for the extraction itself.
Use role evidence to separate a contact from a client opportunity
An HR leader, a hiring manager and a procurement owner can have different parts in the same buying process. Do not assign budget ownership from a title alone. Ask who owns the vacancy, who manages agency relationships and whether there is a preferred-supplier or other procurement route. Store unanswered questions rather than replacing them with a default “decision maker” label.
A person reacting to a job post may be a candidate, colleague or recruiter. That engagement can supply context for inspection; it does not establish hiring authority. The same applies to a reposted vacancy. For a small employer, a founder may own the hiring process, but confirm that responsibility before accepting the profile for a founder-led route.
Prepare a short, evidence-based reason for a conversation: “Your current delivery-team vacancy is in the role family our industrial desk covers. Who handles specialist agency support for that team?” Replace the premise with a real observation. If you mention an available candidate or comparable placement, use a genuine, authorized example. Keep the response and next action with the account so another consultant does not restart the same approach.
Replay the original employer-client CSV exercise
The downloads contain eighteen authored fictional person rows. Every person, employer, vacancy and agency-route observation is invented for teaching. The files are marked synthetic=TRUE and do_not_contact=TRUE; they are not product-export headers, real prospects, client results or an outreach import. Their annotations make the review decisions inspectable.
- Source: all 18 fictional rows.
- CRM reference: three fictional person identities.
- Separate employer-account exclusion reference.
- Expected retained engine output: 15 rows; expected exact person exclusions: three rows; expected identity review: two rows.
- Manual accepted demonstration: six people; four rejected rows; five held rows.
- Manifest: exact counts, original decisions, file hashes and engine version.
Open the local prospect-list workbench in Exclude mode. Load the source as the new audience and the CRM reference as the exclusion file. Select LinkedIn URL for the source key and person_profile for the reference key, using the LinkedIn profile type. Run the comparison. The existing workbench matches the normalized person-profile identities; it does not use names or decide employer fit.
The two files contain 18 source rows and three reference rows. The workbench's Input rows tile shows 21 because it counts both files. Its Main result shows 15 retained rows, Excluded shows three rows and Review shows two rows. R14 has a blank person URL and R15 supplies a company-page URL. Both remain in Main result and also appear in Review. Review is an overlapping warning set, not an extra set to add to the retained count. No reference identity is missing in this particular exercise.
Apply employer exclusions after person-key matching
R11 is the important boundary. It has a valid new person URL, so the person-key engine retains it. Its employer, fictional account A10, is an existing protected client in the separate account reference. A consultant therefore rejects R11 from this new-client delivery. Changing contact title or finding another person at A10 does not create another eligible employer account.
Match account rules with your real, maintained employer identifiers. A company name can be ambiguous across subsidiaries, spelling variants and trading names. The exercise supplies explicit fictional account IDs to make the lesson testable; a real LinkedIn export does not automatically supply your CRM's account IDs. Resolve the actual account mapping before using it to suppress or accept a person.
| Rows | Decision | Reason |
|---|---|---|
| R04 to R09 | Accept: six people at five employers | Fictional current role, employer boundary, hiring context and agency route satisfy the illustrative brief. R04 and R05 belong to the same account. |
| R10 to R13 | Reject: four rows | External staffing supplier, protected current client, wrong territory or candidate engagement without hiring responsibility. |
| R14 to R18 | Hold: five rows | Missing or invalid person identity, unknown supplier route, closed old vacancy or unresolved current hiring remit. |
Use the existing-CRM exclusion guide for incomplete reference files, repeated source profiles and identity repair. Then perform the employer and role review against your own brief. The accepted file in this exercise is a separate manual partition; the workbench did not certify its six rows. Use the agency approval guide when someone else must sign off on the boundary and delivery specification.
Translate the accepted gap into the next extraction job
Track accepted people and employers separately. Six approved people at five companies are six possible role contacts and five employer accounts. If the delivery contract specifies employer accounts, collecting more people at one company will not satisfy it. If it specifies people, count unique reviewed identities and set any per-account rule in the brief.
The illustrative brief requests 250 accepted people. With six already accepted, the remaining accepted-person gap is 244. That is a delivery requirement, not a claim that 244 additional raw profiles will pass review. The eighteen exercise rows were selected to teach boundaries, so six divided by eighteen is not a usable-yield forecast.
For your own desk, start with the published 100 free profile leads at signup, no card, released in batches of 20. Review the real export against the same brief. Count exclusions, repeated identities, role rejections and holds, then record how many unique people are actually accepted. If that real sample is relevant to the next source, use its observed counts as a conditional planning scenario in the one-job extraction planner. An unknown or zero accepted count needs more evidence before a raw-volume estimate.
Once the accepted gap requires more extraction, choose the current commitment for that work. Leadsforlinked's published 500-profile offer is €1.99 and one-time, subject to eligibility and current checkout. Basic publishes unlimited scraping and CSV export at €4.99 monthly or €2.99 effective monthly billed yearly, a €35.88 annual total. Those are extraction offers; they do not promise 500 approved employer contacts, verified emails or messaging features on Basic. The occasional-campaign buying worksheet separates one job from repeated payment commitments.
Make the next audience refresh a distinct employer job
Keep a dated accepted file, employer register, hold file and exclusion reference for each run. Before the next run, update current-client and suppression rules, resolve holds and review whether the original hiring context is still useful. Compare new observations with the previous accepted audience rather than giving each consultant another unreviewed copy of the broad search.
Write the next delivery requirement in the brief: which employer segment is missing, which role within those employers is missing and how many additional accepted people or accounts are needed. A new source, territory or role boundary needs a fresh review. This workflow does not require an autonomous monitor or an assumed safe daily collection quota. LinkedIn's prohibited-software policy restricts scraping and automation and describes account risk; collection and outreach need their own review.
Sources and verification
Sources were read on 13 October 2026. The CSV exercise is an original teaching artifact, independently replayable with the existing workbench. Forum reports identify questions to solve; they are not search-volume measurements or outcome benchmarks.
- LinkedIn Help: Search for people, current filters, conversational-query treatment and approximate results.
- LinkedIn: Sales Navigator advanced filters, optional function, seniority and account-list search context.
- Recruitment agency operator, January 2026, reported client-acquisition difficulties and existing specialization.
- Solo recruitment operator, March 2026, reported weak responses despite contact-data tools.
- Leadsforlinked published offers, dated signup allowance and extraction commitments; account eligibility and checkout were not tested.
- Original employer-client exercise manifest, actual local engine counts and separate manual account and role decisions.
Frequently asked questions
Who should a recruitment agency target on LinkedIn for new clients?
Start with employers that fit your desk, then review current internal HR or talent leaders, relevant hiring managers and founders with actual hiring responsibility. A title, vacancy or job-post reaction alone does not establish buying authority or an agency mandate.
Does excluding CRM profile URLs remove all existing client accounts?
No. The person-key workbench removes matching person identities. A new person at a protected employer can remain. Apply the actual employer-account rules separately and resolve account identity before accepting a new-client prospect.
Are the fifteen retained demonstration rows ready for outreach?
No. Two retained rows also appear in identity Review, and account and role review is still separate. The fictional manual partition accepts six people, rejects four rows and holds five. Every example is synthetic and marked do not contact.
How many additional profiles are needed for the illustrative 250-person job?
The six manually accepted people leave a gap of 244 accepted people. The raw extraction requirement is unknown because the hand-selected teaching records do not establish a future usable yield. Review a relevant real export before estimating its raw-volume requirement.
Do I need Sales Navigator or a paid plan to test this employer audience?
Leadsforlinked publishes a standard LinkedIn route and 100 profile leads free at signup, no card, released in batches of 20. Sales Navigator is optional for that route; keep it if its filters serve your desk. Check the real export and current checkout before buying larger extraction access.