The Google Ads Hub

What is Google Ads and how does the auction work?

Most people meet it the way you probably did: typing "google ads" into a search bar, then landing in a dashboard with more levers than a signal box. Behind that dashboard sits Google Ads, the auction that funds Google itself. Advertisers bid, in real time, for the right to appear beside a search, a video or a map pin, and the scale is not modest: Google's advertising business generated 294.7 billion US dollars in 2025, and Alphabet drew more than 70 per cent of its total revenues from online advertising that year.

Google Ads is Google's advertising platform: an auction in which businesses bid to place ads across Search, YouTube, Maps, Gmail, Discover and millions of partner sites, paying only when someone clicks, watches or takes another action they chose to pay for.

The auction itself runs in the milliseconds after someone searches. Your bid matters, but so do your ad's relevance and expected clickthrough, which is why a sharper ad regularly outranks a fatter wallet. Economists have studied why this model swallowed the industry: peer-reviewed research in the Review of Industrial Organization identifies the fundamental economic difference between online and offline advertising as a substantial reduction in the cost of targeting. You are no longer paying to broadcast at everyone; you are paying to arrive at the exact moment of intent.

The ten google ads campaign types

Each type buys a different kind of attention, and the fox's first question is always which attention your business actually needs.

Two of those rows deserve a footnote before you pick. Performance Max reaches customers across all of Google's channels, including YouTube, Display, Search, Discover, Gmail and Maps, from a single campaign, which makes it powerful and, without clean conversion data, powerfully wrong.

Demand Gen, meanwhile, serves across YouTube (including Shorts), Discover, Gmail, Maps and the Google Display Network, and Google is folding its older Display Ads campaigns into it, so the feeds are where visual budgets are heading.

Reviewed 2026-08-11 by the folkfox bench.

If your product is simple and your budget forgiving, run your first campaign yourself: nothing teaches the auction faster than paying for it. The moment spend reaches the point where a wasted fifth genuinely hurts, or automation starts making choices you cannot explain, tell us what you sell and we will map the route three moves ahead.

What makes Google Ads copy earn the click?

The results page is a queue of near-identical promises, and the searcher gives each one about a second. Your ad does not need to be louder than the others. It needs to be the one that sounds like the answer to the exact question they just typed.

You are not writing one ad, you are writing an inventory

A responsive search ad (RSA) is Google's default search format: you supply a pool of assets and the system assembles a different combination for each auction, learning which pairings perform. You can provide up to 15 headlines and 4 descriptions, and Google's own guidance is to fill as many slots as you can.

The common mistake is writing one good headline fourteen more times. Fifteen paraphrases of "quality plumbing services" give the machine nothing to test. Treat the slots as fifteen different angles: the price angle, the speed angle, the proof angle, the objection-killer, the plain description of what you actually do.

Variety is the raw material. The assembly is Google's job; the angles are yours.

Message matching: echo the query, not your brochure

A search for "emergency boiler repair Leeds" should meet a headline about emergency boiler repair in Leeds, not "Home Services Excellence Since 1998". The closer your language sits to the query, the more the ad reads as the answer rather than an interruption. This is why tightly themed ad groups in search campaigns still beat one bucket of everything.

You do not have to guess what people type. The search terms report shows the actual searches that triggered your ads, and it is the best copywriting brief you will ever get for free. Steal the searcher's vocabulary, and add the irrelevant terms as negatives while you are in there.

Proof beats adjectives, every time

"Trusted", "leading" and "premium" are claims anyone can make, so they carry no information. A number, a timeframe or a named outcome is a claim only you can make, and the reader can feel the difference instantly. The fox's rule: if a competitor could paste your headline over their logo, it is not a headline yet.

Here is a worked example with illustrative numbers. A Leeds plumber runs "Reliable Local Plumber, Call Today" against "Leeds Plumber, On Site Within 90 Minutes, 6,000 Callouts Since 2012" and sees click-through rates of 3.1 per cent versus 5.4 per cent. Same budget, same auctions, and the proof-led ad delivers roughly 74 per cent more visits purely because it swapped adjectives for evidence.

The swaps are usually sitting in your business already:

Your four descriptions do the quieter work. Use them to answer the objection the click is hesitating on: the guarantee, the returns policy, the out-of-hours cover, the thing the searcher is silently worried about. One description that removes a doubt outperforms four that repeat the headline in longer trousers.

Pinning: control has a price

Pinning locks an asset to a fixed position, so your compliance line or brand name always shows. The trade-off is real: every pin shrinks the pool of combinations Google can test, and a fully pinned RSA is just an expensive old-format ad. Pin what must appear for legal or brand reasons, and leave the rest of the inventory free to compete.

A sane middle path is one pinned headline at most, then let performance data argue for more. Draft and grade your fifteen and four in the RSA Studio on this hub before you touch the account; it is easier to be honest about weak headlines when they are not live yet. If you would rather have a second pair of eyes on the whole account, that is what our PPC team does all day, and the door is open.

A fox never announces the hunt. It learns the field, moves once, and makes the move count. Write your fifteen headlines the same way.

What can Google Ads audiences do that keywords cannot?

A keyword tells you what someone wants right now. An audience is a defined group of people you can target or observe based on who they are and what they have done: past visitors, existing customers, active researchers of a category. Keywords read the query; audiences read the person behind it.

This is not a gimmick, it is the economics of the medium. Peer-reviewed research in the Review of Industrial Organization identifies the fundamental economic difference between online and offline advertising as a substantial reduction in the cost of targeting. Audiences are where that cheap targeting lives, and where most accounts leave it switched off.

Remarketing windows: memory with a sell-by date

Remarketing audiences collect people who visited your site or took an action, then hold them for a membership window you choose. The window is the strategy. A visitor from yesterday and a visitor from three months ago are not the same prospect, so do not pay for them as if they were.

Match the window to the buying cycle. Someone abandoning a basket for a 40 pound product might deserve a short, aggressive seven-day audience, while a B2B software evaluation could justify a patient 90-day one (both figures illustrative, and both are choices you make, not defaults you inherit). Stacked windows, say 7, 30 and 90 days, let you bid hardest where the scent is freshest.

First-party lists: the audience you already own

Uploaded customer lists turn your own records into targeting. You can exclude existing buyers from acquisition campaigns, build upsell campaigns for loyalists, or use a list as the seed that teaches Google what a good customer looks like. No keyword can do any of that, because a keyword has no idea who has already paid you.

These lists matter beyond search too. Performance Max treats audiences as signals rather than strict fences, so a strong first-party list quietly steers its machine learning towards people who resemble your best customers.

In-market and custom segments: borrowing Google's memory

In-market segments group people who are actively researching a category, judged from their recent behaviour across Google's properties. Custom segments go further: you build a group from the search terms, sites and apps that define your ideal buyer, including the terms your competitors are known for.

Keywords only see one search at a time. Segments see the pattern behind dozens of them, which is how you reach someone comparing suppliers before they ever type your category into the search box.

Observation versus targeting: watch before you narrow

Every audience you attach to a campaign runs in one of two modes. Targeting mode restricts delivery to that audience only; observation mode changes nothing about who sees your ads, it simply reports how each audience performs and lets you adjust bids for it.

The fox watches the field before it commits. Attach every plausible audience in observation mode, gather a few weeks of data, then promote the proven ones with bid adjustments or dedicated campaigns. Targeting mode on day one is how accounts strangle their own reach on a hunch.

Layering on search: same keyword, different bid

Here is where audiences and keywords stop competing and start compounding, especially across your search campaigns. Suppose your base bid on a competitive keyword is 2.40 in your account currency (illustrative numbers throughout this example). A 14-day basket abandoner might earn a 40 per cent uplift to 3.36, an in-market observer a modest 15 per cent to 2.76, while your past-converter list is excluded entirely because those people have already bought.

One keyword, three different decisions, and not a penny more spent on strangers. That layering logic sits at the heart of any grown-up paid search strategy: the query earns the auction, the audience decides how hard you fight it.

Keywords listen to the question. Audiences remember who asked it last time, and the fox has always trusted the trail it has walked before.

Which Google Ads bidding strategy should you trust with your money?

Bidding is where Google Ads stops being a marketing question and becomes a money question. Choose badly and the auction will cheerfully spend your budget teaching itself things you already knew. Choose well and the machine handles the hour-by-hour arithmetic no human should be doing.

The manual end of the spectrum

Manual CPC is exactly what it sounds like: you set a maximum cost per click for every keyword, and nothing moves unless you move it. It is slow, honest work, and it remains the right choice while your account is new and your conversion tracking is not yet trustworthy. Total control, zero learning.

Maximise clicks is the first rung of automation, with Google spending your budget on the most traffic it can find. That makes it a reasonable data-gathering setting for a brand-new search campaign, and a poor permanent home. Clicks are a cost, not an outcome, and a strategy that optimises for cost alone will find you plenty of it.

What smart bidding actually is

Smart bidding is Google's family of auction-time strategies (maximise conversions, maximise conversion value, target CPA and target ROAS) that sets a different bid for every individual auction, using signals no manual bidder can see: device, hour, location, audience and browsing context among them.

This is not magic, it is economics. Peer-reviewed work in the Review of Industrial Organization identifies the fundamental difference between online and offline advertising as a substantial reduction in the cost of targeting, and smart bidding is that reduction taken to its logical end, priced one auction at a time.

The flavours differ in what they promise. Maximise conversions chases volume within your budget and guarantees nothing about efficiency. Target CPA lets you name a price per conversion, while target ROAS lets you name a return per unit of spend, and that one needs revenue data to function at all.

The prerequisite nobody advertises

Smart bidding learns from conversions, which means it starves without them. Before any automation, your tracking has to be real: measuring web conversions requires giving Google Ads access to data from your website, typically through the Google tag or a linked Analytics property. Automate on top of broken tracking and you have automated a rumour.

Volume matters as much as accuracy. Feed the algorithm three conversions a week and it will swing your bids around on statistical noise, then report the chaos with a straight face. Most people searching for "google ads automation tools" are really asking one smaller question: do I have enough conversion data for the machine to learn from.

A decision path by data volume

Here is the working map we use, counted per campaign over a trailing month. The exact thresholds are folkfox's rules of thumb rather than laws of physics, so treat them as illustrative waypoints, not tripwires.

Count conversions where the bidding happens, not across the whole account. And note that some campaign types remove the choice entirely: Performance Max runs on smart bidding only, so the thresholds above tell you when you are ready for it, not whether you fancy it.

A worked example, on the back of an envelope

Suppose you spend 3,000 a month (in your account currency, and every figure here is illustrative) at an average CPC of 2.50. That buys 1,200 clicks, and at a 2 per cent conversion rate you bank 24 conversions a month. That is maximise conversions territory: enough signal to learn from, too little for a strict target.

Now improve the landing page and lift conversion to 4 per cent, which gives you 48 conversions a month at a blended cost of 62.50 each. You have crossed the threshold, so move to target CPA and set the target near the 62.50 you actually pay, not the 40 you wish you paid. Squeeze the target too hard, too early, and the algorithm simply stops entering auctions, so volume falls and nothing else improves.

Graduation, not loyalty, is the theme. Each strategy is a stage on a path walked at the speed of your data, and pacing that walk is precisely what a good PPC partner earns its keep on. At folkfox we treat every move between strategies as a decision the data has to make first.

The fox lets the machine do the running; it keeps the thinking for itself.

How much should a Google Ads budget be, honestly?

Every agency hears this question, and most dodge it with a rate card. Here is the honest version: your budget is the output of arithmetic you can do this afternoon, not a figure someone hands you. The inputs are your click price, your conversion rate and the number of customers you actually want.

It helps to know what you are buying into. Advertising generated 294.7 billion US dollars for Google in 2025, more than 70 per cent of Alphabet's total revenues. The auction that funds all of that is competitive by design, which is exactly why guessed budgets leak.

The maths that sets the number

Start with three numbers: your average cost per click, your conversion rate, and your monthly conversion target. Divide 100 by your conversion rate to get the clicks one conversion needs. Multiply that by your CPC for the cost of a conversion, then by your target for the monthly budget.

A worked example, with illustrative numbers rather than a forecast. Suppose clicks cost £2.50 and 2.5 per cent of visitors enquire: one conversion needs 40 clicks, so it costs £100. Twenty enquiries a month therefore needs £2,000 a month, or roughly £66 a day.

You do not have to guess the first input either. Google's own Keyword Planner shows estimated monthly search volumes for the keywords you are considering, which anchors the whole calculation in observed demand rather than optimism.

Break-even before a penny leaves

A budget means nothing until you know what a conversion is worth. If your average customer brings £250 in gross profit and a conversion costs £100, the maths smiles at you. If the customer is worth £80, no budget of any size fixes it: the leak is the conversion rate or the click price, not the total.

Run this test before you spend, not after. It takes ten minutes with last quarter's figures and saves months of polite disappointment.

Daily versus monthly pacing

Budget pacing is how Google smooths your spend towards a monthly total rather than metering an identical amount every day. Busy days spend more, quiet days spend less, and the month is what balances. Judge performance on the month, never on one alarming Tuesday.

Set the daily figure from your monthly maths, then leave it alone for a fortnight. Constant fiddling teaches you nothing that patience would not, and it muddies the very data you are paying to collect.

When a budget is too small to learn

There is a floor, and it is not a pound figure: it is data. A budget that buys three clicks a day can take months to produce enough conversions to tell you anything, and automated bidding is even hungrier for signal than you are. Spreading £5 a day across everything is how accounts stay confused for a year.

The fix is concentration, not surrender. One tight campaign, one location, your single best offer can learn on a small budget where five scattered campaigns cannot. That is why we usually start with search campaigns, the most measurable corner of the account, before anything broader.

Do the sums with your own numbers

The Budget Calculator in this hub's Tools den runs the whole chain for you: type in your own click price, conversion rate and target, and it lays out the daily, monthly and break-even figures in plain sentences. Nothing is predicted on your behalf, because nobody should be inventing your CPC. Your numbers in, your budget out.

And if you would rather have a guide who has walked this field before, the PPC team at folkfox does the arithmetic, the structure and the pruning as one job. Bring your margins and we will bring the maths: start the conversation.

The fox never asks how big the forest is. It asks how far the next meal is, then paces itself to arrive first.

Why does Google Ads conversion tracking come before everything?

Conversion tracking is the least glamorous part of Google Ads and the most consequential. Conversion tracking is the mechanism that tells the platform which clicks became something you actually value: a sale, a qualified lead, a booked call. Everything else in the account, the bidding, the budgets, the reports you read on a Monday morning, sits downstream of it.

Google is refreshingly blunt about the prerequisite. To measure conversions on your website, Google Ads needs access to data from your website, typically through the Google tag or a linked Google Analytics property. No tag means no data, and no data means no judgement.

Why does the tag come before the campaign?

Because a campaign launched without tracking spends real money to learn nothing. Smart Bidding steers by conversion signals, so weeks of untracked spend are not a head start, they are an expensive anecdote. You cannot backfill the sales the system never saw.

The fox move is dull and devastatingly effective: install the tag, fire a test conversion, watch it land in the account, and only then write your first ad. For Search campaigns this is important. For Performance Max, which gives you no keywords to steer with, conversion data is the entire steering wheel.

What is the difference between primary and secondary conversions?

Primary conversions feed the Conversions column and, crucially, Smart Bidding. Secondary conversions are observation only: recorded and reported, but never chased. The setting looks like admin and is actually strategy.

Mark everything primary (purchases, newsletter sign-ups, brochure downloads) and the algorithm will cheerfully hunt the cheapest of them, which is rarely the sale. Keep the money events primary and demote the curiosities to secondary. You get a clean signal for the machine and still keep the diagnostic breadcrumbs for yourself.

Do conversion values matter if my numbers are rough?

Yes, because approximate beats absent every time. Without values, every conversion weighs the same, so a £15 enquiry and a £1,500 one look identical to the machine. With even rough values attached, bidding can favour the leads that actually pay your invoices.

A worked example, with illustrative numbers. A heating firm closes one boiler installation from every five quote requests, at roughly £700 gross profit each, so a quote request is worth about £140; a radiator enquiry closes one time in four at £40 profit, so about £10. Feed in 140 and 10 as conversion values and value-based bidding stops treating those two leads as equal cousins.

Your figures do not need an accountant's blessing. Directionally honest values, reviewed each quarter, will outperform valueless counting in every month between reviews.

What breaks silently when tracking is missing or wrong?

The cruel part is that nothing visibly fails. Ads keep serving, clicks keep arriving, and the account looks perfectly alive while quietly steering itself into a hedge.

Every dashboard in the account drinks from this one pipe, and the google ads reporting tools will not warn you when the pipe is cracked. They simply narrate a fictional business with total confidence. This is why a proper audit starts at the tag, not at the ad copy.

So run the boring checklist before the exciting one: tag installed, test conversion received, money events set as primary, values attached even if approximate. If you would rather have a second pair of eyes on the plumbing, the folkfox PPC team reads conversion setups the way a good mechanic listens to an engine, and you can start that conversation here.

The fox does not chase the first rabbit it sees; it learns which runs lead somewhere, then commits. Tag first, spend second.

How do you choose Google Ads keywords that actually pay?

Every keyword is a small wager on what a stranger wants next. Most accounts lose that wager before the auction starts, because the list was built on hunches and headline volume rather than intent. The fix is a method, and it takes an afternoon.

Keywords are the words and phrases you tell Google to match your ads against: each one is a bid on a guess about the searcher's next move. They power search campaigns specifically; shopping campaigns match on product data instead, so this guide is for the words.

Sort every keyword into one of four intent tiers

Buying-intent keywords name the service plus a signal of urgency or place: "emergency plumber near me". These searchers have a problem and a wallet open, so they justify your strongest bids and your sharpest landing page.

Comparison keywords weigh options: "combi boiler vs system boiler". The searcher is close but not committed, which makes them ideal for a helpful guide page and a modest bid, not a hard sell.

Research keywords ask questions: "why is my boiler losing pressure". They are cheap for a reason, because most of these people want an answer, not an invoice. Brand keywords are searches for your own name, worth defending inexpensively if competitors bid on it, and worth leaving alone if nobody does.

Seed the list from your service pages, not a blank page

Open the pages you actually sell with and write down the nouns and verbs your customers would use, not the ones your industry uses. A page that sells "boiler installation" should seed "boiler installation cost", "new boiler fitted" and "boiler replacement", each in its own intent tier.

At folkfox we start every list this way, because a keyword with no page behind it is a click with nowhere to land. If a valuable keyword has no matching page, that is a content brief, not a bid. The same discipline runs through our wider PPC practice: the account should mirror the site, three moves ahead of the searcher.

Why volume alone will mislead you

A big number next to a keyword tells you how many people typed it, not how many meant it. "Boiler" dwarfs "boiler installation manchester" on volume, yet the giant is mostly students, landlords and the idly curious, while the minnow is a homeowner with a cold house.

Real numbers come from Google's Keyword Planner, which shows estimated monthly searches for any keyword you feed it. Read those estimates through the intent tiers above: a hundred buying-intent searches beat ten thousand research ones. Once the list exists, pressure-test the economics with the free google ads tools on this hub before a penny leaves your account.

A worked mini-list for a plumber

Here is how the method lands for an imaginary Manchester plumber. Every figure is illustrative, chosen to show the shape of a good list rather than to predict your market.

Notice the biggest number gets no budget at all. That single decision is where most of the money in this account is saved.

Prune with the search terms report

Your keywords are guesses until real searches test them. The search terms report shows the actual queries that triggered your ads, and any irrelevant ones can be added as negative keywords so you stop paying for people looking for something you do not sell.

Check it weekly for the first month, then monthly. A keyword list is a hedge you trim, not a monument you build once.

The fox does not chase every scent in the wood; it follows the one that ends in dinner.

Why do Google Ads landing pages decide the whole account?

You can build the sharpest campaign imaginable: disciplined keywords, tight ad groups, copy that sings. Then the click lands on a page that loads like treacle, says something slightly different, and demands a phone number before it has earned a hello. Everything you spent upstream decided who arrives; the landing page decides what happens next.

The economics back this up. Peer-reviewed research in the Review of Industrial Organization identifies the fundamental difference between online and offline advertising as a substantial reduction in the cost of targeting. In plain terms, finding the right person is now the cheap part, so the contest has moved to what happens after the click. Targeting only pays when the destination converts.

A landing page is the single page a visitor reaches immediately after clicking your ad, built to continue that ad's exact promise and invite one clear next step. It is not your homepage wearing a lanyard.

Message match: finish the sentence the ad started

A click is a small act of trust. The visitor searched for something, your ad promised it, and the page has about a second to say that promise back to them. If the ad offered "emergency boiler repair in Leeds" and the headline reads "Welcome to our plumbing services", the trust evaporates before the hero image finishes loading.

The fix is almost embarrassingly simple: echo the ad's language in the page's headline, word for word where you can. This matters most in search campaigns, where the visitor arrives holding a very specific question and will judge you on whether you answer it or gesture vaguely towards it.

One page, one promise

A homepage has to serve everyone: investors, job hunters, returning customers, the curious. A landing page serves one visitor with one intent, which is precisely why it converts better. Every extra navigation link, secondary offer or "you might also like" panel is a door out of the room you paid to fill.

The fox's rule: one promise, one action, zero detours. If a page element does not move the visitor towards the single next step, it is working for a different page and should go and live there.

Speed and the mobile reality

Most of your clicks will arrive on a phone, often on a patchy connection, often mid-distraction. A page that feels brisk on your office fibre can feel broken on a train outside Peterborough. Test on a mid-range handset over mobile data, because that is where your budget actually lands.

The pleasant surprise is that this work compounds. The same speed, structure and clarity that convert paid clicks also feed your organic and AI visibility, which is exactly the ground our SEO and GEO services cover. Fix the page once and every channel that touches it gets faster.

Forms that respect the visitor

Every form field is a toll booth. Some tolls are fair: a name and an email are the price of a conversation. But a first-touch form demanding company size, budget range and a direct line reads less like an invitation and more like an interrogation.

Ask only what the next step genuinely needs, and say what happens after submission. "We reply within one working day" costs nothing and answers the visitor's real, unspoken question: is anyone actually behind this page?

The before and after, in numbers

The figures here are illustrative, not a forecast, but the shape of the maths is real. Say an account spends 2,000 a month in its own currency at 2 per click: 1,000 visitors. A slow, mismatched page converting at 2 per cent produces 20 enquiries at a cost of 100 each.

Now fix the message match, cut the form to three fields and get the page loading briskly on mobile. A 4 per cent conversion rate becomes a reasonable ambition: 40 enquiries at 50 each, from the same ads, the same auction and the same budget. Nothing upstream changed; the page simply stopped leaking.

That is why the landing page decides the whole account. Doubling a conversion rate halves your cost per lead in one move, and no bid strategy on earth can match that arithmetic. If you would like a second pair of eyes on yours, tell us where the clicks land and we will follow the trail.

A trail is only as good as the den it leads to, so the fox digs before it hunts.

What do Google Ads match types actually do to your traffic?

A keyword match type is the setting that tells Google how closely someone's search must relate to your keyword before your ad can enter the auction. It is the quietest control in the account and the most consequential. It decides whether your budget meets the searches you planned for, or a long tail of maybes.

The three types, as they behave today

On paper the hierarchy is simple. Broad match casts the widest net, phrase match holds the middle ground, and exact match keeps the tightest grip on intent. In practice, every one of them has loosened over the years, and the names now describe a spectrum of trust rather than a set of rules.

Close variants: why exact is not exact

Every match type now includes close variants: misspellings, plurals, reorderings and searches Google judges to share your keyword's intent. So [emergency plumber] can match "plumber emergency near me", and that is exact match on its best behaviour. The brackets buy you precision of meaning, not precision of wording.

This is not a reason to panic, it is a reason to watch. The looseness usually helps, catching real customers who phrase things oddly. It only hurts when nobody is reading what actually matched, which is a habit problem, not a settings problem.

When broad match earns its keep

Old-school advice said broad match burns money, and on manual bidding it often did. Modern broad match is a different animal: paired with a conversion-based smart bidding strategy, it evaluates each individual search with signals a keyword list cannot see, then bids accordingly. It is the same appetite for machine-found audiences that powers Performance Max, just kept inside your Search campaigns.

The catch is fuel. Smart bidding learns from conversions, so broad match on a campaign with patchy tracking or a handful of conversions a month is a sat nav with no signal. Give it accurate conversion data and a sensible target, and it becomes your prospecting arm while exact match holds the ground you have already proven.

Here is a worked example, with illustrative numbers. Suppose an accountancy firm runs [tax return accountant] on exact match at a 3.00 average CPC and a 5 per cent enquiry rate: 100 clicks cost 300 and produce five enquiries at 60 each. A broad match keyword on the same budget might pull the average CPC down to 1.80 and find 167 clicks, and even at a softer 3 per cent enquiry rate that is five enquiries at exactly the same cost, plus a search terms report full of new phrasings to mine.

That trade only pays while the quality holds. The moment broad match drifts into searches you cannot serve, the maths above collapses, which brings us to the habit that keeps it honest.

The habit that makes any match type safe

The search terms report shows how your ads performed against the actual searches that triggered them, and Google's own guidance is to add irrelevant search terms as negative keywords so your ads stop showing to people looking for something you do not sell. That one report is the difference between match types working for you and working on you.

Make it a weekly ritual, fifteen minutes at most. Sort by spend, add anything irrelevant as a negative, and promote any search that keeps converting into its own exact match keyword. Broad match explores, exact match consolidates, and the negatives fence the territory in between.

Do that for a month and the argument about which match type is "best" quietly dissolves. You will run all three, each doing the job it is built for, on evidence from your own account rather than folklore. If you would rather have a fox reading the trail with you, the folkfox PPC team does this every week for accounts like yours.

The fox does not chase every scent in the wood; it follows the trails that lead somewhere, and it rereads them every morning.

How do negative keywords stop Google Ads waste?

Every Google Ads account leaks. Not through the keywords you chose, but through the searches Google decided were close enough to them. A negative keyword is a word or phrase that blocks your ad from entering the auction when it appears in a search, and it is the cheapest optimisation in paid search because it costs nothing and saves real money.

The search terms report habit

The search terms report shows the actual queries that triggered your ads, not the keywords you bid on, and the two lists are rarely the same. Google's own guidance is blunt about the fix: adding irrelevant search terms as negative keywords prevents your ad showing to people looking for something you do not sell. Most accounts we audit have never opened it.

Make it a weekly ritual: fifteen minutes, sorted by spend, newest terms first. Anything with cost and no conversions gets one of three verdicts: block it, keep watching it, or build a keyword around it. The Waste Finder screen in this hub's Tools den does the sorting for you: export your search terms as a CSV, feed it in, and it flags the likely drains.

Negatives do not match like positives

Here is the trap that catches even experienced hands. Positive keywords stretch: broad match reaches synonyms, related searches and rephrasings on your behalf. Negative keywords do not stretch at all.

A negative does not extend to close variants, so blocking "job" leaves "jobs" live, and blocking "cheap" leaves "cheapest" happily spending. You have to add the plurals, misspellings and variations yourself. The fox checks both directions before crossing: assume every negative blocks only the exact words you gave it, in the match type you gave it, and audit accordingly.

Account level or campaign level?

Structure is what separates a negative keyword list from a pile of panic additions. Two tiers cover almost everyone. An account-level list holds your never-ever terms (jobs, careers, salary, free), applied everywhere at once so you write them exactly once.

Campaign-level lists then handle traffic steering: your emergency-callout campaign takes the planned-project terms as negatives, and vice versa, so campaigns stop bidding against each other. This matters most in search campaigns, where you control the keywords, but it is just as vital where you do not. Shopping campaigns have no positive keywords at all, since ads match on your Merchant Center product data instead of keywords, which makes negatives the only steering wheel you have.

Automated formats like Performance Max give you far fewer levers besides, so a clean account-level list quietly earns its keep across every campaign type you run.

A starter list for service businesses

If you sell a service (trades, clinics, agencies, consultancies), the same intruders turn up in every account. Load these before you spend a pound.

A worked example, with illustrative numbers. A plumber spends 1,200 a month in account currency, and the search terms report shows 18 per cent of clicks came from searches containing "jobs", "course" or "how to fix". That is 216 a month, 2,592 a year, redirected back towards emergency callouts by an afternoon's list building.

Trim the waste first and every other lever pulls harder, because budget freed from junk queries funds the searches that convert. If you would rather a fox ran the sweep for you, our PPC team does this weekly for every account we manage, or you can ask us to audit yours.

The fox does not chase every rustle in the hedgerow; it learns which ones are the wind, and saves its legs for the real thing.

Why does Quality Score change what you pay per click?

You are never billed your bid. You are billed just enough to beat the advertiser beneath you, and how much counts as enough depends on how Google rates your ad's quality. That rating is the quiet lever most accounts never pull.

Quality Score is Google's 1 to 10 diagnostic of how useful your keyword, ad and landing page are to the person searching. It is not a dial you can buy; it is a report card on relevance. The live auction weighs the same underlying signals, so the report card shows you exactly where money is leaking.

What are the three components?

Expected clickthrough rate asks whether people who see your ad tend to click it, given its position. Ad relevance asks whether your copy actually answers the query behind the keyword. Landing page experience asks whether the page after the click is fast, honest and useful.

Each component is graded below average, average or above average, which makes triage refreshingly simple. Find the one sitting below average, fix that first, and leave the rest alone for now.

How does the auction discount relevance?

Every search triggers an instant auction, and your position is set by Ad Rank: broadly, your bid multiplied by your quality signals, adjusted by context. A relevant ad earns its slot with less money, while an irrelevant one has to pay its way in. The discount is structural, not a loyalty perk.

Economists treat these auctions as one of the defining features of the medium. Avi Goldfarb's overview in the Review of Industrial Organization argues that the fundamental economic difference between online and offline advertising is a substantial reduction in the cost of targeting, and names auctions among the field's main research themes. Quality Score is that targeting economics made visible in your account.

Here is the arithmetic, with illustrative numbers. Suppose a rival bids £4.00 with weak quality signals worth a multiplier of 5, for an ad rank of 20, while you bid £2.60 with strong signals worth 9, for 23.4. You sit above them yet pay only enough to clear their 20, which works out at about £2.23 rather than your full £2.60.

That is the discount in plain sight. Relevance is not rewarded out of kindness; better ads keep searchers clicking, and clicking keeps the whole machine earning.

Which fixes are worth your effort?

Start with an hour's work. Split any ad group where one headline is serving twenty keywords, and pause keywords whose intent your page cannot honour. This alone often lifts expected CTR, because the ad finally speaks to one searcher instead of a crowd.

Next, book an afternoon. Rewrite your ads so the query's language appears in a headline and the description promises precisely what the page delivers. This is the core discipline of well-run Search campaigns, and it moves expected CTR and ad relevance together.

The big job is landing page experience, which is a build task rather than a settings task. Speed, mobile layout, message match and honest content all count, and the same page work compounds in organic and AI search, which is where folkfox's SEO and GEO work overlaps with paid. If the page is the weak link, no amount of ad polish will out-write it.

One warning deserves its own line: never chase the score for its own sake. A 10 on a keyword that never converts is a beautifully polished loss, so read Quality Score beside your conversion data. If the whole account needs a second pair of eyes, that is what our PPC team is for.

The fox seldom outbids anyone; it knows the ground so well the auction charges it less for crossing.