Managing the Most Expensive Clicks in Google Ads

In most accounts, a bad click costs a few pesos and you learn from it in aggregate. In legal, insurance, medical devices, restoration, and a handful of other categories, one bad click can cost more than a small business spends on ads in a day.

That single fact changes the job. Everything that's "hygiene" in a normal account becomes the entire discipline, and several habits that work fine at low CPCs become expensive at high ones.

TL;DR

Query control is the job. At high CPCs, one irrelevant click can cost more than a week of optimisation is worth.

Broad match needs a much higher bar — it's a different risk at $80 a click than at $0.80.

Geography and scheduling are bid levers, not settings you configure once.

Check search terms daily at first, not weekly.

What changes at high CPC

One wasted click is a real loss

At $0.50, a hundred junk clicks is $50 and a shrug. At $80, ten junk clicks is $800 — a meaningful share of a monthly budget gone on queries that were never going to convert.

This is why negative keyword work stops being maintenance and becomes the primary daily activity for the first few weeks of any high-CPC account.

The learning period is genuinely expensive

Smart Bidding needs conversions to learn. At $80 a click and a 5% conversion rate, thirty conversions costs roughly $48,000 of spend.

Practical consequences:

Don't restructure casually. Every rebuild restarts learning at a price you can feel.

Start on manual or a conservative target and move to fuller automation once there's data — the reverse of what I'd advise on a cheap ecommerce account.

Don't chop and change bid strategies. Each switch costs another learning period.

Small percentage improvements are large absolute ones

A 10% improvement in conversion rate on a $50,000/month legal account is worth more than most entire small accounts. That justifies effort that would be over-engineering elsewhere: bespoke landing pages per matter type, call-recording review, proper offline conversion plumbing.

The controls that matter, in order

1. Query control

Check the search terms report daily for the first two to four weeks of any new high-CPC campaign, then weekly. Not because it's virtuous, but because a single un-negated term can burn thousands before a weekly review catches it.

Build negatives in layers:

Account level: jobs, careers, salary, free, DIY, "how to", "what is", student, template, meaning.

Campaign level: other practice areas or services you don't want.

Ad group level: terms that belong in a sibling ad group.

Legal accounts specifically bleed on: "free", "pro bono", "how much do lawyers charge", "can I sue", "law school", competitor names you didn't intend, and — critically — the *other* side of the matter, where a firm acting for one party pays for clicks from the opposing one.

2. Match type discipline

Broad match plus Smart Bidding works well in many accounts. At $80 a click, the same setup can spend thousands exploring queries before it learns.

A workable sequence:

Start with phrase and exact. Learn what actually converts at a controlled cost.

Introduce broad later, in a separate campaign with its own budget cap, once you have conversion data worth matching against.

Never run broad without a mature negative list.

3. Geography, treated as a bid lever

For location-bound services, geographic precision is one of the strongest efficiency controls you have. Not just "which cities" — which suburbs and postcodes have historically produced signed matters or booked jobs.

Also: set location targeting to presence, not interest, unless you deliberately want people merely searching about a place. The default catches a lot of expensive irrelevance.

4. Scheduling against answering capacity

If a $100 click generates a call nobody picks up, you paid $100 for a voicemail.

Look at conversion rate by hour and by day, then stop buying the hours you can't answer. Same argument as emergency trades campaigns, with a bigger price tag attached.

5. Landing page relevance

Quality Score affects what you pay per click. At high CPCs a modest Quality Score improvement is worth real money on every click, every day — the mechanics are in Quality Score, and the speed side in landing page speed.

One page per matter type or service. A generic homepage receiving $80 clicks is an expensive way to be irrelevant.

Habits to unlearn

"Set broad match and let the algorithm figure it out." Fine at $1. Costly at $80 before a negative list exists.

"Test lots of things quickly." Each test costs a learning period. Fewer, better-reasoned changes.

"Weekly search terms review." Too slow at the start.

"Judge on cost per lead." The whole point of high-CPC categories is that lead value varies enormously — see cost per signed case.

What actually justifies the price

High CPCs aren't irrational. Competitors bid that much because the matters or jobs are worth it. The question is never "is this click expensive" — it's "is this click expensive relative to what it's worth to me."

Which is why the economics come first. Without a defensible maximum cost per acquisition, you're either underbidding into invisibility or overbidding into losses, and you can't tell which.

Want the controls audited?

Most high-CPC accounts I audit are leaking on two or three query patterns that a proper negative structure would have caught, and buying hours nobody covers. Both are same-week fixes with immediate effect.

That's part of my law firm PPC and PPC strategy work. Send the account and your target cost per acquisition through the project fit page.

Related reading:

Cost Per Signed Case

Negative Keywords: The Lever Most Accounts Ignore

Google Ads Quality Score

The Weekly PPC Management Routine