Boolean Search for Recruiters: 3 Strings, One Method

One string cannot both size a market and source from it. Mercedes Gordo, who spent two years at LinkedIn in Dublin before running delivery at Nova, on the three-string method and the syntax notes nobody hands you.

11-minute read

Key takeaways

Write three strings, not one. Broad to size the market, balanced to actually source from, precise for when the balanced one returns noise. Two is enough for a simple role.
Synonyms in the local language are the biggest coverage multiplier. A search in Berlin that misses Backend-Entwickler, or one in Madrid that misses Planificador de Demanda, loses profiles no amount of tuning gets back.
LinkedIn does not support wildcards. No asterisks, no brackets, no braces. Parentheses are the only grouping symbol it recognizes, and AND, OR and NOT have to be uppercase or they count as search terms.
Add exclusions one at a time. One exclusion too many hides good profiles and gives you no signal that it happened. Check the result count after each one.

Why one Boolean string is never enough

The usual failure is not bad syntax. It is asking a single string to do two incompatible jobs. A string wide enough to tell you how many people in this market could do the job is, by definition, too wide to source from. A string tight enough to give you a clean list is too tight to tell you whether the list is small because the market is small or because your criteria are wrong.

So you write more than one, deliberately, and each one answers a different question. That is the whole method, and it takes about ten minutes once the criteria are decided.

Before any of it: decide the criteria first. The real job titles and their variants, not the pretty internal one. The location and whether a radius works. The experience band. The must-have skills, separated from the nice-to-haves. The languages and the actual level the job needs. And who to exclude. A string built before those are settled just encodes the confusion faster.

The three strings, side by side

String	What it is for	What it catches	What it misses
Broad	Sizing the market before you commit	Titles and their synonyms, nothing else. Maximum coverage	Nothing much. That is the point, and why it is unusable for outreach
Balanced	The one you actually source from	Titles plus one or two must-have skills, with basic exclusions	People who describe the job differently or do not list the tool on their profile
Precise	When the balanced one returns noise	Adds seniority, more skills and firmer exclusions. Few results, highly qualified	Strong profiles with an unusual title or a thin profile write-up

Two strings are enough for a straightforward role. Use three when there is a nuance that justifies the middle step: a specific tool, a seniority band, or an industry that matters.

String 1: broad, to size the market

Titles and synonyms joined with OR, in quotes so each one matches as a phrase, and nothing else. No skills, no location filter inside the string, no exclusions.

Broad

("Backend Engineer" OR "Back-end Engineer" OR "Backend Developer" OR "Software Engineer"
 OR "Backend-Entwickler")

Catches everyone who calls the job by any of its names. Unusable for outreach, and that is not what it is for.

Do this: run the broad string first and read the result count, not the profiles. That number is your market size, and it tells you whether the role is hard because there are few people or hard because the criteria are wrong. Those two problems have opposite solutions.

String 2: balanced, the one you source from

Take the titles group and join it with AND to a second OR group holding one or two must-have skills. Then a short NOT group for the obvious exclusions. This is the string you will spend the day in.

Balanced (recommended)

("Backend Engineer" OR "Backend Developer" OR "Backend-Entwickler" OR "Software Engineer")
AND (Go OR Golang OR Kotlin OR "distributed systems")
NOT (intern OR internship OR Praktikum OR recruiter)

Catches backend engineers who list at least one of the core technologies, minus interns and other recruiters. Misses anyone who does not name the tool on their profile, which is more people than you would expect.

Notice the exclusion in German. Praktikum is doing the same job as intern, and leaving it out means every intern in the market stays in your results. The same applies in reverse: exclusions written only in the local language miss profiles kept in English.

String 3: precise, when the balanced one is noisy

Add the seniority into the titles, stack a second skills group, and widen the exclusions. You will get few results and most of them will be right. Use it when the balanced string returns more than you can read, not as a starting point.

Precise

("Senior Backend Engineer" OR "Staff Engineer" OR "Lead Backend Developer"
 OR "Senior Software Engineer")
AND (Go OR Golang)
AND (Kubernetes OR Kafka OR "event-driven")
NOT (intern OR Praktikum OR freelance OR consultant OR recruiter)

Catches senior backend engineers with the specific stack. Misses strong people whose title is unusual, which is why you never run this one on its own.

Or skip the strings entirely.
Describe the profile once and Nova Recruiter sources it. Free tier, no credit card.

Try for free

The two things that decide whether a string works

Synonyms, in the local language and in English

This is the largest single source of missed coverage, and it is invisible: you never see the profiles a bad synonym list excluded. Every title and every skill needs its variants in the language of the market you are searching and in English, because people write their profiles in both.

A demand planning search in Spain that only looks for "Demand Planner" misses "Planificador de Demanda" and "Supply Planner". A backend search in Germany that only looks for "Backend Engineer" misses "Backend-Entwickler". Include the abbreviations and the alternative role names too.

Do this: write the synonym list before you write the string, in both languages, and keep it. It is the reusable part: the operators change per role, the synonyms for a job family barely change at all.

Exclusions, added one at a time

Exclusions feel productive because the result count drops, which is exactly why they get over-used. The usual set is interns, freelancers, consultants and other recruiters, plus profiles openly advertising availability if you are after passive candidates. Beyond that, every extra NOT is a bet that nobody good uses that word.

Do this: add exclusions one at a time and check the count after each. If a single word cuts the list in half, it was not filtering noise, it was filtering people.

The syntax rules LinkedIn documents and almost nobody mentions

These are not tips, they are in LinkedIn's own help page on Boolean search, and every one of them silently changes what your string returns if you get it wrong.

The operators have to be uppercase. LinkedIn's wording is that AND, NOT and OR must be typed in uppercase for the search to work correctly. A lowercase and is read as a search term, so you are quietly searching for the word itself.
Parentheses are the only grouping symbol LinkedIn recognizes. Braces, square brackets and angle brackets are not supported. If you brought a string over from another tool that uses them, it is not doing what it looks like it is doing.
There are no wildcards. An asterisk does not truncate a word, so engineer* does not catch engineering. You have to write the variants out, which is another reason the synonym list matters so much.
+ and - are not officially supported, even though they sometimes appear to work. Use AND and NOT.
Stop words are ignored inside quotes. LinkedIn drops words like by, in and with from quoted phrases for performance, so an exact phrase built around one of them is not as exact as it looks.

One correction worth making, because it circulates as fact: Boolean does work in the main LinkedIn search bar, not only in Recruiter or Sales Navigator. What those two add is the field-level filtering around the string, not the operators themselves.

And if you have no seat at all, use the Google X-ray variant. Same logic, wrapped so Google only returns profiles:

Google X-ray

site:linkedin.com/in ("Backend Engineer" OR "Backend Developer") AND (Golang) "Berlin"

Costs nothing. The trade-off is real: you only reach profiles Google has indexed, and you lose every filter the platform would have given you.

One more thing worth knowing if you are working inside Recruiter Lite: the operators work, but the tier caps how far you can reach, and our breakdown of Recruiter Lite's limits covers what that means in practice.

Get the string written for you

This whole method is one of eleven free AI skills our team packaged up. You give it the role and the market, it asks which platform you search in, and it returns the calibrated strings with the synonyms in the right languages, the exclusions and a line on what each one catches and misses. It runs on Claude, ChatGPT or Gemini.

If your platform is Nova Recruiter instead, it returns the search as structured filters rather than a Boolean string, because there is nothing to write: the criteria go in directly. The companion skill for what to send the people you find is in our guide to recruiting message templates.

Get the Boolean skill, free.
Eleven AI skills for recruiters. Works with Claude, ChatGPT and Gemini.

Get the skills

What this looks like without Boolean at all

Boolean is a way of describing a person to a keyword index. It works, and it will keep working, but the effort goes into translating a clear idea into operators and then tuning the operators until the list looks right.

Nova Recruiter takes the description instead. You say what the profile is and the agent searches 800M+ public profiles, ranked by Talent Score rather than keyword density, which is the part a Boolean string genuinely cannot do: a string can tell you who mentions Golang, not who is good. Where that sits in the wider picture is in our breakdown of the three levels of sourcing automation.

The bottom line

Boolean is not hard, it is just usually done in one pass when it needs three. Size the market broad, source from the balanced string, tighten only when the noise makes you. Get the synonyms right in both languages, and treat every exclusion as a bet you might be losing. That is most of the craft, and it holds whether you type operators or describe the profile in a sentence.

Frequently asked questions

What is Boolean search in recruiting?

Combining search terms with the operators AND, OR and NOT so a search platform returns exactly the profiles you want. In practice it means grouping job titles and their synonyms with OR, joining them to required skills with AND, and cutting out the profiles you do not want with NOT. Quotation marks force an exact phrase, so "Product Owner" does not match a profile that happens to contain both words separately.

Does Boolean search work in the normal LinkedIn search bar?

Yes. LinkedIn documents AND, OR and NOT as working in the main search bar, not only in Recruiter or Sales Navigator. What those two add is the field-level filtering around the string. The rules that catch people out are LinkedIn's own: the operators have to be uppercase, parentheses are the only grouping symbol it recognizes, there are no wildcards, and stop words like "in" and "with" are dropped from quoted phrases.

How many Boolean strings should you write per role?

Two for a simple role, three when there are nuances worth a middle step: specific skills, a seniority band, or an industry that matters. One string is the mistake, because a single string has to be either broad enough to size the market or precise enough to source from, and it cannot be both at once.

What should you exclude from a Boolean search?

The usual suspects are interns, freelancers, consultants and other recruiters, plus profiles openly advertising availability if you are sourcing passive candidates. Exclude in the local language too: in Germany that means Praktikum as well as intern. And add them one at a time. One exclusion too many hides good profiles, and you will never see which ones you lost.

How do you do a Boolean search on Google?

It is called an X-ray search, and it is the fallback when you have no Recruiter seat. You wrap the same logic in a site restriction so Google only returns LinkedIn profiles: site:linkedin.com/in ("Backend Engineer" OR "Backend Developer") AND (Golang) "Berlin". You lose the filters a recruiting platform gives you, and you get a searchable index that costs nothing.

Do you still need Boolean if you use an AI sourcing tool?

For the search itself, no. The reason to learn it anyway is that writing a good string forces you to decide what you are actually looking for: the real titles, the must-have skills, the exclusions. That thinking is the same whether you type it as operators or describe it in a sentence, and a vague brief produces a bad shortlist either way.

About the author

Mercedes Gordo

Head of Delivery and Customer Success

Mercedes Gordo is Head of Delivery and Customer Success at Nova. Mercedes spent two years at LinkedIn in Dublin as a relationship manager and account director, and previously ran Customer Alliance's Spanish operation.

Connect on LinkedIn →

Stop tuning strings.
Describe the profile once and let Nova Recruiter source it across 800M+ profiles.

Try for free Book a demo