Boolean Search and Sourcing: Practical Workflow for Recruiters

Boolean search is the technical foundation of recruiter sourcing. The same logical operators that power web search — AND, OR, NOT, parentheses, quotation marks for exact phrases, wildcards — drive the candidate-discovery query inside LinkedIn Recruiter, Google X-ray searches against LinkedIn public profiles, GitHub user search, ATS keyword search, and specialized sourcing tools like SeekOut, Hiretual, and RecruitEm. A sourcer who can write a precise Boolean query finds the qualified candidate population an order of magnitude faster than one who relies on default keyword search.

This article walks through the practical Boolean-search workflow: the core operators, the common query patterns for different role types, the platform-specific syntax variations, the iterative refinement process that distinguishes effective sourcers from struggling ones, and the limits of Boolean search where modern semantic-search tools take over.

Data Notice: Search-platform syntax shifts over time as platforms update their query parsers. The patterns documented here reflect platform behavior at time of writing; verify current syntax in platform help docs before launching a high-stakes search. The recruiter tooling evaluation page covers the broader sourcing-platform ecosystem.

The core Boolean operators

Five operators do the bulk of sourcing work:

AND narrows results — every term joined by AND must appear. Example: python AND django AND "machine learning" returns only profiles mentioning all three terms.
OR broadens results — any of the OR-joined terms satisfies the query. Example: python OR ruby OR golang returns profiles mentioning any of the three. OR is often combined with parentheses for grouped alternatives.
NOT excludes results — terms after NOT are filtered out. Example: python AND NOT recruiter filters profiles where Python appears but the person is a Python recruiter rather than a Python engineer.
Parentheses group operators for precedence control. Example: (python OR ruby) AND (django OR rails) AND NOT recruiter reads as “(python or ruby) developers using (django or rails), excluding recruiters.”
“Quoted phrases” match the exact phrase rather than individual terms. Example: "machine learning engineer" matches the exact title rather than profiles that mention machine and learning and engineer separately.

A sixth operator, the wildcard *, is supported on some platforms (Google) and not others (LinkedIn). Where supported, it expands single-term variants — data* matches data, database, datasets.

The composition of these operators determines query selectivity. A query that’s too broad returns thousands of profiles to triage; a query that’s too narrow returns the candidate the sourcer already knew about. The skill is in hitting the precision/recall sweet spot for the role at hand.

Query patterns by role type

Different role types have different Boolean signatures.

Engineering roles typically use a stack-based query pattern: required language(s), required framework(s), required infrastructure or domain, with experience-level proxies in the title.

("Senior Software Engineer" OR "Staff Software Engineer")
AND (Python OR "Go") AND (Kubernetes OR Docker)
AND NOT (recruiter OR "talent acquisition")

Sales roles typically use a quota-and-segment pattern: title proxies for senior IC vs. management, market segment proxies, and product-category proxies.

("Account Executive" OR "Senior Account Executive")
AND (SaaS OR "Software-as-a-Service") AND ("mid-market"
OR "enterprise") AND NOT recruiter

Product roles typically use a function-and-domain pattern:

("Senior Product Manager" OR "Principal Product Manager")
AND (B2B OR enterprise) AND (analytics OR data OR ML)

Senior leadership roles often use a trajectory pattern, chaining title patterns over a career arc rather than a single-snapshot query. These queries often run as a sequence of refinements rather than a single Boolean expression.

The pattern that distinguishes strong from weak sourcers isn’t just operator fluency — it’s the ability to map a role’s actual requirements (the job description) to a query (the search expression) that selects for those requirements without selecting for irrelevant signals.

Platform-specific syntax

Each major sourcing platform has subtle syntax differences worth knowing:

LinkedIn Recruiter supports AND, OR, NOT (must be capitalized), parentheses, and quoted phrases. Wildcards are not supported. Title field, company field, location field, and skills field accept Boolean expressions separately.
Google X-ray search supports site:linkedin.com/in/ to constrain to public LinkedIn profiles, full Boolean syntax including wildcards, and Google-specific operators like intitle:, inurl:, and date-range filters. Google X-ray is the workaround when LinkedIn Recruiter access is unavailable.
GitHub user search supports its own operator syntax: language:Python, location:"San Francisco", followers:>100, repos:>5. The Boolean concepts carry over but the operators look different.
ATS keyword search typically supports basic Boolean but the indexing is often weak; many sourcers use ATS-side search for talent rediscovery and external platforms for fresh sourcing.

Translating a query from one platform to another is part of the sourcing workflow. A query that works on LinkedIn Recruiter often needs syntactic adaptation for Google X-ray or GitHub.

Effective Boolean sourcing is iterative, not one-shot. The typical workflow:

Write the initial query based on the job description’s stated requirements. Run it. Inspect the result count and a sample of returned profiles.
If the count is too high (~5,000+ profiles): add AND clauses for more specific requirements. Add NOT clauses for common false positives (recruiter, intern, freelancer if the role is full-time employment).
If the count is too low (~10 or fewer profiles): relax requirements. Convert AND to OR for tier-2 nice-to-haves. Drop overly specific tool requirements in favor of broader category terms.
Inspect the top 20 profiles that come back. Are they on-target? If not, what false-positive pattern is showing up? Add a NOT clause to filter it. What true-positives are missing? What are they listed under that the query didn’t catch?
Iterate until the result set is high-precision. Then begin outreach.

Sourcers who skip the iteration phase and start outreach from a noisy result set burn outbound capacity on mismatched candidates. The 30 minutes spent refining a query saves hours of bad outreach downstream.

For more on the candidate-engagement side that follows sourcing, see passive candidate engagement.

When Boolean isn’t enough

Modern sourcing tools layer semantic search and AI-driven candidate-matching on top of Boolean. SeekOut, Hiretual, and similar tools take a job description and return ranked candidates without requiring the sourcer to write a Boolean query at all. These tools handle the precision/recall tradeoff via embedding-based similarity rather than logical operators.

Boolean still matters because:

The semantic-search systems are opaque — when they return bad results, the sourcer needs Boolean to investigate why and to construct the targeted query that fixes the gap.
Niche populations (researchers in a specific subfield, engineers with specific compliance certifications) are often best found with hand-tuned Boolean against a professional graph rather than via embedding similarity.
The cost structure of semantic-search tools makes them expensive at high volume; Boolean against LinkedIn Recruiter and Google X-ray is the cost-effective default for most sourcing.

The mature sourcing function uses both: Boolean for precision-driven niche searches, semantic-search tools for broad-stroke matching at high volume.

For broader sourcing strategy context, see sourcing channel effectiveness and talent pool and pipeline strategy.

Common Boolean mistakes and how to fix them

Several recurring mistakes show up in sourcer queries that look reasonable on the surface but underperform in practice. First, treating OR clauses as soft preferences rather than hard alternatives. Sourcers often write (python OR ruby OR golang OR java OR javascript) thinking it broadens the search; in practice it broadens to anyone who lists any of the five languages, including profiles that mention them in passing or in unrelated roles. The fix is to use OR sparingly for true alternatives (python or ruby for a polyglot role) and use the platform’s primary-language signal where available. Second, over-relying on title fields when skills are the target. A query like ("Senior Software Engineer") filters to people with that exact title, missing the substantial population working as Staff Engineer, Tech Lead, Software Architect, or Principal Engineer who do equivalent work. The fix is to query the skills field and the company-prestige proxy in addition to title, accepting more triage volume in exchange for broader recall. Third, NOT clauses that exclude too aggressively. A query that says NOT (recruiter OR manager OR director OR VP) to filter out non-IC roles will also filter out engineers who happened to mention “I worked with my engineering manager on…” in their summary. The fix is to constrain the NOT to the title field specifically, not free-text. Fourth, not iterating on returned results. Sourcers who write a query, run it, and immediately start outreach miss the precision-tuning step that distinguishes high-response sourcing from low-response sourcing. Fifth, neglecting platform-specific quirks — LinkedIn requires capitalized AND/OR/NOT, Google X-ray accepts lowercase, GitHub uses different operator syntax entirely. Queries copy-pasted across platforms without adaptation often silently fail to parse, returning results based only on the keyword matches without the operator logic.

Takeaway

Boolean search is the technical foundation of recruiter sourcing. Five operators (AND, OR, NOT, parentheses, quoted phrases) compose into queries that select candidate populations matching specific role requirements. Strong sourcers iterate queries against returned results to hit the precision/recall sweet spot before launching outreach. Platform-specific syntax variations matter; modern semantic- search tools layer on top of Boolean rather than replacing it. The discipline is mapping the role’s actual requirements to a query that selects for those requirements without catching irrelevant signal.

Sources

Schmidt, F. L., & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262–274.
Sackett, P. R., & Lievens, F. (2008). Personnel selection. Annual Review of Psychology, 59, 419–450.
Boolean Black Belt (Glen Cathey). (Ongoing). Boolean search techniques and sourcing methodology resources.
Sourcing Institute. (Ongoing). Sourcing certification curriculum and Boolean technique documentation.
LinkedIn Talent Solutions. (Recurring). Recruiter search best practices and platform syntax documentation.
Aptitude Research. (Recurring). Sourcing tools and technology effectiveness studies.