Job Seekers Clicking Google Ads? Block These Terms

Adding "jobs" as a negative keyword catches about a third of job seekers

The standard advice — add jobs, careers, hiring, vacancy to a negative list and move on — works on the terms where the person spelled out what they wanted. Those are the minority. Most people looking for work in your industry never type the word job.

Here is a real spread of the kind of terms a plumbing company pays for, none of which contain "job", "career" or "vacancy":

• plumber salary uk
• how much do plumbers earn per hour
• how to become a plumber at 30
• plumbing apprenticeship near me
• gas safe registration course
• trainee plumber no experience
• plumber cv example
• is plumbing a good trade
• indeed plumber manchester

Every one of those is somebody who wants to be a plumber, not hire one. A broad or phrase keyword like plumber near me or a Performance Max campaign with a plumbing feed will happily match several of them. And the person clicking is motivated — they will read the whole page, find the phone number, and call. Job seekers are not bounce-and-leave traffic. They are click, read, call traffic, which is exactly what makes them expensive.

The fix is not a longer list of job words. It is recognising that non-commercial intent comes in three families, each with its own vocabulary, and that the vocabulary is specific to your industry.

Family one: job seekers, and the four clusters they search in

Job-seeker traffic clusters around four things: pay, training, platforms and your brand. Cover all four and you catch most of it.

The pay cluster

salary, wage, wages, earn, earnings, pay, pay rate, hourly rate, how much do, starting salary, average salary, salary uk, per hour.

This is the biggest cluster in trades, care, logistics, hospitality and construction. hgv driver pay, care assistant hourly rate, how much do electricians make, warehouse operative salary.

One caution: hourly rate and call out charge are ambiguous. A homeowner searching emergency plumber call out charge london is a buyer. A school leaver searching plumber hourly rate uk is not. If you block hourly rate as a phrase negative you lose both. In practice salary, wage and earn are clean, rate and cost are not — block the clean ones and read the rest.

The training cluster

apprenticeship, apprentice, course, courses, training, qualification, qualifications, certification, nvq, diploma, city and guilds, how to become, work experience, internship, graduate scheme, entry level, no experience.

This one is under-blocked because the words look harmless next to your service. electrician training course sits one word away from electrician and nobody flags it until it has been running for six months. If you sell to businesses, add the certification acronyms your industry uses — they are almost always training searches, never buying searches.

The platform cluster

Job-board brand names get typed alongside the trade constantly: indeed, reed, totaljobs, cv library, glassdoor, linkedin, monster, seek, jobsite, ziprecruiter. Also agency in some markets, which is genuinely ambiguous, and recruitment, recruiter, staffing, which are not.

The brand cluster

[your brand] careers, [your brand] jobs, [your brand] vacancies, [your brand] application form, working at [your brand], [your brand] reviews employee. If you run a brand campaign — and if you do, you are bidding on your own name with broad-ish coverage — this traffic lands in it. Brand campaigns get the least scrutiny in the account because the CPA looks good on average. The job-seeker terms hide behind that average.

In service accounts we audit — trades, care, logistics, hospitality — job-seeker terms usually account for somewhere between a low single-digit percentage and around a tenth of clicks, and the range is wide enough that the only useful thing to do is look at your own report rather than trust a benchmark. Outside those industries it is often close to zero. A dental practice gets very little of it. A national care provider gets a lot.

Family two: students, researchers and the professionally curious

The second family clicks because they want to understand something, not buy it. They cost less per click than job seekers, on average, and they appear in far greater numbers, especially in B2B and technical accounts.

The markers are consistent across every industry:

pdf, ppt, pptx, doc, slideshare, wikipedia, meaning, definition, define, what is, what are, types of, list of, examples, example of, advantages and disadvantages, pros and cons of, diagram, formula, equation, explained, notes, lecture, assignment, essay, thesis, dissertation, mcq, quiz, exam, syllabus, history of, who invented.

For an industrial pump distributor that produces search terms like centrifugal pump working principle, types of pumps pdf, npsh formula, what is cavitation in pumps, pump diagram labelled. For a CRM vendor: crm meaning, crm advantages and disadvantages, types of crm systems pdf, crm case study. For a law firm: what is probate, power of attorney explained, intestacy rules diagram.

The phrasing advantages and disadvantages of X deserves its own mention. It is an assignment title. Nobody who is about to spend money phrases a search that way. Same with case study of, history of and anything ending in pdf.

The part that is genuinely hard

what is is not automatically worthless. If you sell something with a long consideration cycle and you have a real top-of-funnel offer — a guide, a calculator, a webinar — then what is asset finance is a lead you can work. If your only conversion action is a purchase or a booked appointment, those terms almost never reach it.

So the test is not "is this term informational". The test is "does my account have anything to sell this person". Most accounts do not, and pretend they do. If you have been running for six months and the whole what is cluster has produced no conversions, you have your answer. If you are unsure whether you are looking at a targeting problem or a conversion-tracking problem, work through the diagnosis in order before you start cutting.

Family three: freebie hunters, DIY and second-hand

The third family wants what you sell but has no intention of paying for it. Three sub-groups:

Free-seekers. free, free download, free version, free trial (sometimes), free alternative, open source alternative, freeware, no cost, without paying, crack, cracked, nulled, torrent, keygen, serial, mod apk. Software and digital-product accounts get hit hardest. [competitor] free alternative and is there a free version of [your product] are both very common and both a poor use of a click if you have no free tier.

DIY-ers. how to fix, how to repair, how to install, myself, diy, step by step, tutorial, youtube. A boiler repair company pays for how to fix a boiler pressure drop myself all day. That person will call you only if the fix fails, which is a lottery ticket, not a marketing channel.

Bargain hunters. cheap, cheapest, free delivery, discount code, promo code, voucher code, coupon, second hand, used, pre owned, refurbished, ex display, gumtree, ebay, facebook marketplace, craigslist, rental, hire, borrow.

The word "free" is the single most dangerous negative in this article

If you add free as a broad negative you will block free quote, free consultation, free survey, free site visit, free delivery over $50, free trial and free plan. For a huge number of advertisers, those are the best-converting terms in the account. Adding one three-letter negative can remove a third of a lead-gen account's volume overnight, and because it happens silently, people spend a fortnight blaming Google.

Block the phrase, never the word: free download, free version, free alternative, free template, free pdf, free course, free crack. Phrase negatives only block terms containing that exact word sequence, which is exactly what you want here. This is one of several places where negative match types behave differently from positive ones, and where broad negatives quietly do more than you asked.

Every word on these lists is a buyer term in somebody else's account

This is why copy-pasting a "1,500 universal negative keywords" list from a blog post is a bad idea, and why we are giving you vocabulary clusters rather than a file to import.

• salary is a buyer term for payroll software, HR consultancies, salary benchmarking tools and recruitment agencies.
• cv and resume are buyer terms for CV-writing services and career coaches.
• apprenticeship, course, training, nvq are buyer terms for every training provider and college in the country.
• free is a buyer term for any freemium product and any service that opens with a free quote.
• diy and how to install are buyer terms for a builders' merchant selling the parts.
• second hand, used, refurbished are buyer terms for a refurb reseller.
• pdf and template are buyer terms for document software.
• indeed and linkedin are buyer terms for anyone selling recruitment tooling or competing with them.

Five minutes reading a proposed list against your own offer prevents the expensive mistakes. It also keeps the list short, which matters more than people think — see how many negative keywords are too many for why a bloated list creates its own problems.

Where this traffic hides: Performance Max, Shopping and the report itself

Three structural reasons this traffic survives longer than it should.

Performance Max. Search terms for Search, Shopping and Performance Max all report in campaign_search_term_view, so you can see PMax search terms — but PMax will not accept shared negative keyword lists. Negatives go on the campaign, one campaign at a time, up to 10,000 of them. If you run twelve PMax campaigns, your job-seeker vocabulary has to be added twelve times and maintained twelve times. That friction is why most PMax campaigns have a handful of negatives and a Search campaign in the same account has four hundred. The 10,000 cap and how to spend it is worth understanding before you start bulk-adding.

Shopping. Standard Shopping takes negative keywords and shared lists like Search does, but matching is driven by your product feed, not by keywords you chose. If a product title contains the word "free" — free-standing bath, gluten free flour — you cannot block free without breaking the campaign. Shopping negatives work differently and need feed-aware choices.

The report. The search terms report does not show every term you paid for. Low-volume terms are withheld, which means a long tail of one-click job-seeker and student queries never appears as a line you can act on. You will see the recurring ones and pay silently for the rest. What the report hides, and roughly how much of your spend can sit behind that wall, is covered in the search terms report and what it hides. Broad-match keywords and PMax asset groups are also structurally prone to reaching this kind of query in the first place, for the reasons set out in why Google Ads shows irrelevant search terms.

The conversion trap: when a job application counts as a lead

Here is the failure mode that turns a small leak into a self-reinforcing one.

Your website has one contact form. It is your only conversion action. A job seeker lands on the site from care assistant salary near me, cannot find a careers page, and uses the contact form to ask whether you are hiring. Google Ads records a conversion. Smart Bidding now has evidence that care assistant salary near me converts, and bids accordingly. Two months later that query family is one of your highest-spending segments, with a CPA that looks acceptable on the dashboard and a real-world value of zero.

The same thing happens with phone calls tracked as conversions. Call assets in service accounts get dialled by applicants constantly, and a call over 60 seconds is counted as a conversion whether it was a booking or a job enquiry.

This also limits what any automated tool can do for you, including ours. PPCleaner never blocks a search term that converted in the last 90 days. That rule exists to make the product safe to leave running. It also means that if your conversion tracking counts job applications as leads, we will treat those terms as valuable and leave them alone. No automation can see through a conversion action that is lying to it.

The fix is on your site, not in the ads account:

1. Give applicants somewhere to go. A careers page linked in the main navigation removes most of the misdirected form fills.
2. Split the conversion actions. A separate careers form, not counted as a conversion, or counted as secondary.
3. If you use call tracking, exclude or separately label the number on the careers page.
4. Go back 90 days in the search terms report and check which "converting" terms contain salary, apprentice, cv or vacancy. Those conversions were applications.

None of this stops the click. You have already paid for it by the time anyone fills in anything. It stops the bidding algorithm from buying more of them.

A one-sitting cleanup you can do this afternoon

Thirty to sixty minutes, depending on account size.

1. Open the search terms report. Date range: last 90 days. All campaigns, including PMax.
2. Add the Cost, Clicks and Conversions columns, and Campaign.
3. Filter the search term field by contains for one marker word at a time. Start with: salary, wage, earn, apprentice, course, training, cv, resume, vacanc, hiring, recruit, indeed, career, internship.
4. Then the student markers: pdf, meaning, definition, what is, types of, advantages, examples, diagram, wikipedia, essay.
5. Then the freebie markers: free download, free version, free alternative, crack, torrent, diy, how to fix, second hand, used, coupon, voucher.
6. For each filter, read the terms before adding. You are looking for the ones that are unambiguous in your business.
7. Add as phrase negatives. Put the ones that are safe account-wide into a shared negative list applied to your Search and Shopping campaigns. Add the same terms at campaign level in each PMax campaign, because lists do not apply there.
8. Note the total cost of what you just blocked. That is your monthly recovery, roughly, and it is the same exercise as the broader hunt described in how to find wasted ad spend in Google Ads.

Then put a reminder in the calendar for four weeks' time, because the vocabulary regenerates. New job-board brands appear, seasonal training searches spike in September and January, and a single new asset group in PMax will open a fresh set of matches.

What automation changes here, and what it does not

The manual version of this works. Its problem is that it is a batch job done by a human who has eleven other accounts, so it happens in month one and then in month seven.

PPCleaner reads the search terms your account paid for every night across Search, Shopping and Performance Max, judges which ones were never going to convert, and adds them as negative keywords — including the PMax campaign-level adds that nobody does by hand twelve times. It does not touch budgets, bids, ads or keywords. Every block is reversible in one click for 90 days, which is the realistic answer to "what if it blocks something I wanted": you look at the log and undo it.

What it cannot do, stated plainly:

• It cannot see the terms Google withholds from the report. Nothing can.
• It will not block a term that converted in the last 90 days, so a contact form that counts job applications as leads will shield job-seeker terms from it indefinitely. Fix the tracking first.
• It cannot tell that salary is a money word in a payroll-software account without your account's own conversion history telling it so. Terms that convert stay.
• It reports nothing for Display, Video or App campaigns, because those campaign types report no search terms at all.

The job-seeker leak is worth fixing by hand once, today, so you see the size of it. After that the question is only whether you want to do it again every month.