Google Ads for Tradespeople: Emergency Jobs That Pay

Google Ads works for tradespeople on emergency jobs and usually loses money on planned ones. Someone searching emergency electrician near me at 21:40 has already decided to buy and will take the first firm that answers the phone. Someone searching rewire cost is six weeks from a decision and will get three quotes anyway. So the honest rule for Google Ads for tradespeople and emergency jobs: bid hard on urgent intent, inside a radius you can actually reach, only during hours a human answers, with a tight negative keyword list, and only if your gross profit per job is at least three times your cost per booked job. If one of those four fails, switch the campaign off. It is arithmetic, not marketing.

Why 2026 makes this harder, not easier

Click prices on urgent trade searches keep climbing, because national lead-generation firms and franchise networks bid against you on the same fifteen words and can afford a loss on the first job. Meanwhile automated campaign types hide the detail: broad match, Smart Bidding and Performance Max report a healthy-looking cost per conversion while the search terms underneath show you paid for electrician apprenticeship, how to wire a socket, and a town ninety minutes away.

Automation is not the enemy. Automation without a manual control list is. A one-van firm spending a few hundred a month cannot afford the learning period a machine-learning bidder wants, and does not produce enough conversions to feed it. Small budgets need guardrails.

Google Ads for tradespeople: emergency jobs versus planned work

The expensive click is often the cheap job. A high emergency click price with a one-in-two close rate beats a cheap planned-work click with a one-in-eight close rate. Never judge a keyword on click price. Judge it on cost per booked job.

One exception: high-value planned work with a narrow search, such as a three-phase supply upgrade or commercial fire alarm servicing, can pay, because the order value absorbs several lost quotes. You only find that out by tracking quotes through to signature.

The only maths that matters: click to real order value

Work backwards from money you bank, not from clicks. The figures below are illustrative. Replace them with your own.

• Cost per click on emergency terms: assume £8. It varies wildly by town and trade.
• Click-to-contact rate: a decent mobile landing page for urgent work gets 20 to 30 per cent. Assume 25 per cent. A contact costs £32.
• Contact-to-booked-job rate: with a human answering, 50 to 70 per cent. Assume 60 per cent. A booked job costs about £53.
• Gross profit per emergency job, invoice minus materials minus labour cost, before overhead: say a £320 callout leaves £180.
• Result: £180 of gross profit against £53 of acquisition. Roughly 3.4 to 1. That works.

Now break one link. If nobody answers after 18:00 and half your clicks land in the evening, your click-to-contact rate is 12 per cent, a booked job costs £110, and the ratio falls under 2 to 1, before overhead and before your own time. The campaign is now a hobby.

Set the target before you launch: cost per booked job must not exceed one third of gross profit per job. Write it on the wall. It is the only kill switch you need.

The number most firms get wrong

Google reports conversions, meaning calls and form fills. It has no idea which became a job, and it optimises towards whatever is cheapest to produce, usually tyre-kickers. Once a month, count how many booked jobs came from ads and divide spend by that. A spreadsheet is enough. The gap between reported cost per conversion and true cost per booked job is routinely two or three times.

Radius: draw it around drive time, not distance

• Start at the distance you would genuinely drive at 22:00. For most vans that is 20 to 30 minutes, not 30 miles.
• Draw the radius from where the van actually sleeps, not from your registered office address.
• Set location targeting to presence, not presence or interest. The default shows your emergency ad to someone in another country reading about your town.
• Exclude the postcodes you resent: the congestion zone, the estate with no parking, the island with a ferry. Add them as location exclusions on day one.
• If you cover a city plus a rural fringe, run separate campaigns with separate budgets. Otherwise cheap rural clicks eat the budget and the city jobs never show.

Ad schedule: advertise only when you can answer

An emergency ad running while your phone is on voicemail is a donation. Be honest about answering capacity, then match the schedule to it.

• Map your answer coverage hour by hour for one normal week. Not intentions. What happened.
• Turn ads off in the hours nobody answers. A 24/7 emergency ad with a 09:00 to 17:00 phone is the commonest way a trade account burns money.
• If you do work out of hours, bid up, not down. Fewer competitors are live at 23:00 and the job value is highest. This is the one time an aggressive bid is defensible.
• Pause on days you are already full. A booked-solid Tuesday does not need leads, it needs the phone left alone.

This is where a shared diary earns its keep. If scheduling, job cards and your intake form live in one system — OdinTask does this, and so do others — you can see whether next week has capacity worth advertising for. Advertising into a full diary is how you end up cancelling on a customer who paid emergency rates.

Negative keywords: the list to paste in on day one

Add these as campaign-level negatives before your first click, grouped so you can maintain them.

• Employment: jobs, vacancy, hiring, apprentice, apprenticeship, salary, wage, careers, recruitment, cv. They do not want your callout, they want your payroll.
• Training: course, courses, training, qualification, nvq, city and guilds, college, diploma, exam, certification, how to become.
• DIY and information: how to, diy, tutorial, youtube, wiring diagram, fix myself, reddit, forum, what is.
• Free and cheap: free, cheapest, cheap, bargain, discount, grant, funded, second hand, used.
• Price research: price list, how much, hourly rate, quote calculator, estimate calculator, comparison, vs, versus, reviews.
• Wrong buyer: supplier, wholesale, trade counter, parts, spares, tool hire, insurance, warranty, complaint, building regs.
• Wrong trade or wrong place: every neighbouring town you do not serve, every adjacent trade you do not do.

The price group needs judgement. On planned work, price searches are tyre-kickers. On emergency work, emergency electrician call out charge is a buying signal, not research. Never blanket-block a word across the account. Block it in the campaign where it hurts.

Then do the only maintenance that matters: open the search terms report weekly for two months, monthly after that. Read every term you paid for. Add negatives. Fifteen minutes a week beats any account audit you can buy.

The manual control checklist for automated campaign types

If you run Performance Max or broad match with Smart Bidding on a small budget, insist on these.

1. Account-level and campaign-level negatives, both. Some automated formats respect only some lists.
2. Brand exclusions, so you do not pay to be shown to people already searching for you by name.
3. Conversion defined as booked jobs, or at minimum calls over 60 seconds. A 9-second call is a wrong number, not a conversion.
4. Manual or capped bidding until you have roughly 30 real conversions in 30 days. Below that the algorithm is guessing and you are paying tuition.
5. Read the search terms and asset reports monthly. If a campaign type will not show you what you bought, that tells you something about the campaign type.

When to switch it off entirely

Advertising is not a moral obligation. Turn it off when:

• You are booked four weeks out and turning work away. Spend it on a second van, or on nothing.
• Your work is 90 per cent repeat and referral, and the ads only produce the customers you like least.
• Nobody answers the phone reliably. Fix that first. It is free and it doubles every other number in the chain.
• Cost per booked job has sat above one third of gross profit for two months running, after you tightened radius, schedule and negatives.

A note for readers in Sweden

The maths is identical, but two local details change the answer. ROT relief cuts the customer-facing price of household labour by 30 per cent of the labour cost, up to an annual per-person ceiling shared with RUT — check the current amount at Skatteverket. That makes planned domestic work convert better than in a market with no relief, so a planned-work campaign clears the bar more often. Show the net price in the ad and on the landing page, not the gross. Second, emergency search volume in a smaller Swedish town can be a handful of searches a month. Low volume is not failure, but you cannot judge a campaign in two weeks on it.

The 30-minute setup that beats most agency accounts

1. Write your average gross profit per emergency job and divide by three. That is your maximum cost per booked job.
2. One campaign, emergency terms only, phrase and exact match, no broad.
3. Radius set to a realistic night-time drive, presence only, bad postcodes excluded.
4. Ad schedule matching real answer coverage, bids raised out of hours.
5. Negative keyword list pasted in, all seven groups.
6. Landing page: phone number first, tap to call, one line on response time, no gallery, no essay.
7. Conversion set to calls over 60 seconds. Weekly 15-minute slot to read the search terms report.
8. Review at 30 days against your one number. Above it, cut. Below it, raise the budget.

Everything else is decoration. The firms that make ads pay are not the ones with clever copy. They answer the phone, drive a sane distance, and know what a job is worth before they bid on it. If tracking that chain from click to paid invoice is currently three apps and a notebook, that is the real problem: enquiry forms, the diary, the quote and the invoice belong in one place. Start a free 14-day trial of OdinTask, or see what is included on the pricing page.