Google Ads Budget Waste: A UK Business Guide to Cutting Wasted PPC Spend in 2026

Google Ads wasted spend is the single most recoverable line item in most UK SME marketing budgets. It is not a strategy problem and it is rarely a creative problem — it is an accounting problem hiding inside a reporting interface that, by default, shows you what worked and quietly buries what did not. A business spending £2,000 a month on search is typically sending somewhere between a fifth and a third of that to clicks that could never have converted: wrong search terms, wrong locations, wrong hours, wrong devices, or conversions that were counted twice and taught the bidding algorithm to chase the wrong thing.

This guide leads with diagnosis rather than tactics, because you cannot cut waste you cannot see. It walks through how to find where PPC budget is leaking in a live account, then works through the four levers that consistently close the gap for UK SMEs — negative keyword strategy, quality score improvement, conversion tracking setup that reflects real revenue, and campaign structure changes that give the bidding system clean signals. The aim throughout is a lower cost per acquisition on the same budget, not a smaller budget. Cutting spend is easy and usually costs you volume; cutting waste keeps the volume and removes the cost.

What counts as wasted Google Ads spend?

Wasted spend is any click you paid for that had no realistic path to a sale, plus any click you paid too much for because the system was optimising against a broken or misleading signal. That second half is the part most businesses miss. A click on the search term “free CRM software” when you sell a £12,000 implementation is obvious waste. A perfectly relevant click that cost £9.40 instead of £5.10 because your conversion action fires on every page view, so Smart Bidding believes almost every visitor converts, is invisible waste — and it is usually the larger number.

It helps to separate waste into three tiers. Tier one is targeting waste: search terms, placements, locations, times and devices that should never have been bought. This is the easiest to find and the fastest to fix, because Google shows you the raw data in the search terms report if you go and look at it. Tier two is signal waste: conversion tracking that double-counts, counts the wrong events, or attributes value it cannot substantiate, causing automated bidding to overpay for the wrong audience. Tier three is structural waste: account architecture that forces unrelated intents to share a budget and a bid strategy, so your strongest performers subsidise your weakest without anyone noticing.

The tiers compound. A poorly structured account makes the search terms report harder to read, which delays negative keyword work, which pollutes the conversion data, which degrades the bidding, which raises cost per acquisition across everything. That is why an audit that only adds negative keywords produces a short-lived improvement — the leak reappears within a quarter because the mechanism that created it was never addressed. The sequence in this guide is deliberate: find the leaks, fix the signal, then restructure so the leaks do not refill.

Google Ads waste in UK SMEs — the numbers that matter

The figures below reflect what a first-pass audit typically surfaces in UK SME accounts in the £1,000–£10,000 per month spend band — the range where an account is large enough to matter commercially but rarely large enough to justify a full-time specialist. They are not ceilings. Accounts that have run on default settings since launch, or that changed hands between agencies without a proper handover, sit well above these numbers.

The 62-day figure deserves attention because it is the one that costs the most and gets the least scrutiny. Conversion tracking breaks silently — a website rebuild removes a tag, a form vendor changes its thank-you page URL, a consent banner is reconfigured and starts blocking the tag for a third of visitors. Nothing errors. The account keeps spending. Smart Bidding, deprived of conversion signals, either throttles delivery on your best campaigns or starts bidding on a stale model. By the time someone asks why leads are down, two months of budget has gone. If your website is under active development, treat conversion tracking as a release-gate item in the same way you would treat Core Web Vitals and page performance: something to verify after every deployment, not once a year.

Where the money actually leaks — waste by source

When you decompose recovered waste across a set of audited SME accounts, the distribution is consistent enough to plan around. Irrelevant search terms dominate, but they are not the whole story, and an audit that stops there leaves most of the money on the table. The chart below shows the approximate share of total identified waste attributable to each source.

Irrelevant search terms arrive mainly through broad match and, increasingly, through Performance Max, where the search terms visible to you are a summarised subset rather than the full picture. The classic UK SME pattern is a service business matching on job-seeker queries (“IT support jobs Manchester”), research queries (“what is SD-WAN”), DIY queries (“how to fix…”), competitor names you have no offer against, and free-intent queries (“free”, “cheap”, “template”, “salary”, “course”). Each of these is a negative keyword you can add once and benefit from indefinitely.

Geographic targeting drift is the quietest of the group. The default location setting is presence or interest — meaning Google will show your ads to people outside your targeted area who it believes are interested in it. For a national e-commerce brand that is often fine. For a Reading-based managed service provider who only travels within the Thames Valley, it means paying for clicks from Glasgow, Dublin and, depending on how the campaign was built, from outside the UK entirely. Switching to presence-only targeting is a two-click change that regularly removes a tenth of spend without touching a single keyword.

Search Partners and Display expansion is the setting most often left on by accident. A search campaign created in the standard flow may include Search Partners and, historically, display expansion. Traffic from these networks carries different intent and usually a materially different conversion rate. It is not automatically waste — some accounts genuinely perform on Search Partners — but it should be a measured decision with a segment-level report behind it, not a default nobody examined.

What a wasted-spend audit costs, and what it typically recovers

The economics of PPC waste reduction are unusual: the work is front-loaded, the saving is recurring, and the recurring saving is usually a multiple of the one-off cost. The table below sets out the UK market shape for management and audit work at each spend band, along with the range of waste a competent first-pass audit tends to identify. Figures are exclusive of VAT; Google Ads itself is billed to UK advertisers with UK VAT at 20%, recoverable in the normal way if you are VAT-registered.

Two things are worth reading out of that table. First, the audit cost is broadly flat while the recoverable waste scales with spend, which is why the case for a structured audit strengthens sharply above roughly £2,000 per month. Second, the implementation window is longer than most people expect, and deliberately so. Negative keywords can be added in an afternoon, but conversion tracking rebuilds, campaign restructures and bid strategy changes all need a learning period before the data is trustworthy again — typically two to three weeks per significant change, and they should not be stacked simultaneously or you lose the ability to attribute the improvement.

Beware the percentage-of-spend management model when your primary objective is waste reduction. It creates a structural tension: the fee falls when the budget falls, so there is no financial incentive to recommend spending less on a channel or a campaign. It is not that percentage models are wrong — they align well when growth is the goal — but during a remediation phase, a fixed fee or a project fee removes the conflict. The same reasoning applies more broadly when you are deciding where to place specialist capability in-house versus with a provider: match the commercial model to the outcome you actually want.

Broad match with Smart Bidding versus tightly controlled structure

Most waste arguments eventually reduce to one question: how much control should you hand to Google’s automation? The honest answer depends almost entirely on your conversion volume and the quality of your conversion signal. Automation is a statistical system; starve it of data and it performs poorly, feed it bad data and it performs confidently and expensively in the wrong direction. The two approaches below are both legitimate, and the deciding factor is rarely ideology.

The practical answer for a business under roughly thirty conversions a month is a hybrid: run your known-intent, revenue-carrying terms in tightly matched campaigns with a manual or target-CPA strategy, and run a separate, capped discovery campaign on broad match whose only job is to surface new search terms for you to promote or negate. That discovery campaign should be treated as a research budget with an explicit monthly ceiling — typically 10% to 15% of total spend — and reviewed weekly. It gives you the query discovery benefit of automation without letting it take over the accounts that pay the bills.

One caveat that matters more each year: Performance Max complicates this picture because it spans Search, Shopping, Display, YouTube, Discover and Gmail inventory under one budget, and its reporting does not decompose cleanly into those channels. If you run Performance Max alongside search campaigns, apply account-level negative keyword lists, use brand exclusions deliberately, and check the asset group and search-category reports rather than assuming the headline cost per acquisition tells the truth. A Performance Max campaign that appears to convert cheaply is frequently harvesting brand searches that would have converted anyway.

Account health scoring — where most UK SME accounts sit today

Before you can prioritise, you need an honest read on the account’s current state. The grid below is the diagnostic we work through, grouped into the three tiers introduced earlier. Score each row for your own account: high risk means it is actively costing money now, medium means it is degrading performance, low means it is a refinement rather than a repair.

The pattern in almost every audit is the same: businesses invest attention in the refinement rows — ad copy variants, extensions, ad strength indicators — because those are the things the Google Ads interface actively prompts them about, while the high-risk rows sit untouched for months because nothing in the interface flags them. Ad strength is a useful signal, but it is a measure of asset variety, not of profitability. An account can hold an excellent ad strength rating on every ad group and still be sending a third of its budget to search terms with no purchase intent.

The 90-day waste-reduction timeline — what a real remediation looks like

Waste reduction fails most often not because the diagnosis was wrong but because everything was changed at once. Google’s bidding systems re-enter a learning phase after significant changes to conversion actions, bid strategies or campaign structure, and overlapping learning periods make it impossible to tell which change helped. The sequence below spreads the work across a quarter, ordering it so that measurement is fixed before anything that depends on measurement is touched.

The single most valuable feature of this sequence is the observation window in weeks four and five. It feels like inactivity and it is the opposite: it is the period in which you find out whether your measurement repair actually worked, before you build anything on top of it. Businesses that compress the whole programme into a fortnight typically end up unable to explain their own results, and when performance dips two months later they have no way to identify which of eleven simultaneous changes caused it.

Benchmarks and KPIs — what a healthy UK SME account looks like

Benchmarks are directional, not targets. Cost per click in UK legal or insurance search can exceed £25, while a niche trade service might pay under £1.50 for the same volume of intent, so absolute figures travel poorly between sectors. What does travel is the structural ratios below — they describe account hygiene rather than sector economics, and an account that scores well on these is very unlikely to be leaking badly.

Two of these deserve elaboration because they sit outside the ads account and are therefore routinely excluded from PPC reviews. Mobile landing page speed is a paid media metric whether or not your agency treats it as one: more than half of UK search traffic on most SME accounts is mobile, and a page that takes five seconds to become interactive loses a meaningful share of the clicks you already bought. If you have not measured this recently, the field data in Search Console and a run through PageSpeed Insights will tell you in minutes — the mechanics are covered in our guide to Core Web Vitals and conversion rate.

Response time to enquiries is the other. Paying £7 a click to generate a form fill that sits unanswered until the following afternoon is waste in every sense that matters, even though no PPC report will ever show it as such. In competitive UK service categories the buyer has usually contacted three suppliers; the ordering of the replies matters as much as their content. Before spending money reducing cost per lead, it is worth confirming that the leads you already buy are being handled.

Measurement integrity — the number behind most invisible waste

If you only fix one thing, fix conversion tracking. It is the input to every automated decision Google makes on your behalf, and it is broken or misleading in a clear majority of the SME accounts we review. The most common single defect is a conversion action that fires on a page that any visitor can reach, which inflates the apparent conversion rate, hides the real cost per enquiry, and teaches Smart Bidding to buy traffic that browses rather than traffic that buys.

The repair is methodical rather than technical. Open the conversions summary and list every action with its category, count method, attribution model and primary or secondary designation. For each one, ask a single question: does this event correspond to something the business would recognise as a genuine outcome? Newsletter sign-ups, PDF downloads, contact page views and time-on-site thresholds are all useful diagnostics and none of them should be a primary conversion driving bids. Move them to secondary so they still report but no longer steer spend.

Then check the count method. “Every” is correct for e-commerce transactions where a repeat purchase is genuine incremental revenue; “one” is correct for lead-generation forms, where a prospect submitting three times is one enquiry, not three. Getting this backwards on a lead-gen account is one of the quietest and most expensive errors available, because it inflates conversion counts precisely on the campaigns generating repeat submissions from confused visitors.

Finally, verify the tag actually fires for real users. Open the site in a fresh browser session, interact with the consent banner exactly as a typical visitor would, complete the form, and confirm the conversion appears. UK GDPR and PECR require consent for non-essential cookies, and consent mode is how you preserve modelled measurement while honouring that choice — but a misconfigured banner that blocks the tag unconditionally will silently remove a large share of your conversion data. Businesses running a formal governance programme usually catch this in a data review; those without one tend to find it months later, which is one of several reasons a growing number of UK SMEs bring in senior technology oversight on a fractional basis rather than leaving measurement to whoever built the site.

Negative keyword strategy — building the list that keeps working

A negative keyword strategy is not a list of words you dislike; it is a structured, layered defence with three levels, each doing a different job. Treated as an occasional clean-up task it produces diminishing returns. Treated as an architecture it becomes the single most durable cost control in the account.

Level one is the account-level list, applied to every campaign. This holds the universal disqualifiers — terms that could never represent a buyer for anything you sell. For a typical UK B2B service business that means employment intent (“jobs”, “salary”, “vacancy”, “career”, “apprenticeship”, “recruitment”), education intent (“course”, “training”, “certification”, “tutorial”, “pdf”, “definition”, “meaning”), free intent (“free”, “freeware”, “open source”, “template”, “diy”), and geography you cannot serve. Build this once, review it quarterly, and apply it to every new campaign at creation. In most accounts it is between 150 and 400 terms and it is the highest-leverage hour of work available.

Level two is the campaign-level list, which enforces separation between campaigns that would otherwise cannibalise each other. If you run one campaign for “cloud migration” and another for “managed IT support”, each needs negatives that push the other campaign’s core terms away, so the query lands where you intended and the budget you allocated by intent is actually spent by intent. This is also where brand negatives belong: excluding your brand terms from non-brand campaigns is the only reliable way to see what new-customer acquisition genuinely costs, because brand traffic converts cheaply and will otherwise flatter every campaign it touches.

Level three is the ad-group-level list, which is where you protect margin. If your “emergency” service commands a higher price than your standard offering, negatives at ad group level keep the emergency ad in front of emergency queries and stop the cheaper standard ad from absorbing the click. This layer is finer-grained work and it is worth doing only after the first two are stable.

On match types for negatives, one detail causes recurring trouble: negative keywords do not behave like positive ones. A negative broad match keyword only blocks queries containing all of its words in any order, and negative match types do not match close variants — misspellings, plurals and singulars are treated as separate terms. If you negate “job”, you have not negated “jobs”. Build the obvious variants explicitly, and re-check the search terms report a week later to confirm the block actually took effect rather than assuming it did.

Finally, keep an eye on the limits and the tooling. Account-level negative lists can be shared across campaigns and are far easier to maintain than duplicating terms campaign by campaign; a shared list edited once propagates everywhere it is applied. If you are managing more than a handful of campaigns, put the master list in a shared document with a short note on why each term was added — six months later, nobody remembers why “audit” was excluded, and someone will remove it.

Quality Score improvement — paying less for the same position

Quality Score is Google’s 1–10 diagnostic of how relevant your keyword, ad and landing page are to the searcher, and it feeds directly into Ad Rank — which means it feeds directly into what you pay per click. Two advertisers can occupy the same position at materially different prices. It is not a vanity metric and it is not a ranking factor to be gamed; it is a reasonably honest report card on whether the three parts of your offer agree with each other.

It decomposes into three published components, and the diagnosis is more useful than the headline number. Expected click-through rate asks whether people click your ad when they see it for this keyword. Ad relevance asks whether the ad text actually addresses the query. Landing page experience asks whether the page delivers what the ad promised, quickly and without friction. Add the three status columns to your keyword view, sort by cost descending, and look only at the top twenty spenders. A “below average” on any component for a high-spend keyword is a specific, actionable defect.

The fixes map cleanly onto the components. Below-average ad relevance almost always means the ad group is too broad — twenty loosely related keywords sharing three generic ads. Split it so each ad group covers one tight theme and the ad can repeat the searcher’s language verbatim. Below-average expected click-through rate usually means the offer is undifferentiated: add the concrete detail that makes you the obvious choice for that query, whether that is response time, coverage area, accreditation or price transparency. Below-average landing page experience means the page is slow, generic, or asks the visitor to hunt for what the ad promised.

The economics are worth stating plainly. Moving a set of high-spend keywords from a Quality Score of 4 to 7 typically reduces cost per click somewhere in the region of 15% to 30% at the same position — the exact effect depends on the auction and the competing advertisers, so treat any specific multiplier you read online with caution. What is reliable is the direction: better relevance costs less. That is why quality work and waste work belong in the same programme. Negative keywords stop you buying the wrong clicks; Quality Score work reduces what you pay for the right ones.

One structural note. Landing page experience is assessed on the page, not the site, which means a well-built dedicated page can carry a strong score even on an otherwise mediocre website — and a beautifully designed site can score poorly if every ad points at the homepage. If your ads all land on a single generic page, that is usually the cheapest Quality Score improvement available, and it overlaps almost entirely with the work you would do to improve organic performance for the same terms, as covered in our guide to local search visibility for UK businesses.

The 12-point wasted-spend audit checklist

Work through these in order. The sequence is deliberate — each step either produces evidence the next one needs, or removes a distortion that would make the next one misleading. Set aside roughly three to four hours for a first pass on a single-country account with fewer than ten campaigns.

1. Export 90 days of search terms with cost, clicks, conversions and conversion value. Sort by cost descending, filtered to zero conversions. This is your evidence base and your before-picture; save it somewhere outside the ads account.
2. Inventory every conversion action. Record category, count method, attribution setting and primary or secondary status. Flag anything that fires on a page a visitor can reach without taking a deliberate action.
3. Test the tag end to end as a real visitor. Fresh browser, interact with the consent banner as a typical user would, submit the form, confirm the conversion is recorded. Repeat on mobile.
4. Reconcile reported conversions against CRM records for the same period. A variance beyond roughly 10% means one of the two systems is wrong, and you need to know which before you optimise against either.
5. Check location settings on every campaign. Confirm presence-only targeting unless there is a documented reason not to, and review the user locations report for spend outside your service area.
6. Review network settings. Segment by network and decide on Search Partners and display expansion with performance data in front of you rather than by default.
7. Build or refresh the account-level negative keyword list from the week one export, covering employment, education, free intent, unservable geography and irrelevant competitors.
8. Segment by hour and day of week. Compare when clicks occur with when enquiries are actually answered, and schedule accordingly. Out-of-hours spend is only waste if nobody responds.
9. Segment by device. Compare mobile, desktop and tablet conversion rates and cost per acquisition; investigate before adjusting, because a poor mobile conversion rate is usually a landing page defect rather than a targeting one.
10. Review Quality Score components on your twenty highest-spend keywords and record which component is below average for each, so the remedial work is targeted rather than general.
11. Map every ad group to its landing page and check that the page repeats the ad’s specific promise above the fold, loads quickly on a mobile connection, and asks for exactly one action.
12. Separate brand from non-brand in both campaign structure and reporting, so the true cost of acquiring a new customer is visible rather than averaged with people who already knew your name.

Decision framework — scoring your account’s efficiency readiness

Score one point for each of the twelve checklist items you can honestly answer yes to today, then multiply by eight. That gives a rough efficiency readiness figure comparable to the benchmark below, which represents the median score of UK SME accounts arriving for a first audit. It is not a precise instrument — it is a way of deciding whether your next hour is better spent on ad copy or on measurement.

Interpretation is straightforward. Below 40 means measurement is almost certainly unreliable and no optimisation work should begin until it is repaired — anything you conclude from the current data has a good chance of being wrong. 40 to 70 is the range most SME accounts occupy: the fundamentals are partly in place, the fastest wins are in negative keywords, location settings and conversion action hygiene, and a structured 90-day programme is the right response. Above 70 means the obvious waste has already been removed, and further improvement comes from harder work — offline conversion import, landing page testing, margin-aware bidding and audience layering — with correspondingly smaller individual gains.

The score also tells you something about where to put the money. An account below 40 does not need a bigger budget or a better agency retainer; it needs a fortnight of unglamorous diagnostic work, most of which an in-house marketer can do with the exports and this checklist. An account above 70 is where specialist time genuinely earns its fee, because the remaining gains require judgement rather than hygiene.

Common mistakes that keep budgets leaking

The failure patterns below turn up repeatedly, and they share a characteristic: each one looks like sensible practice from inside the Google Ads interface. That is what makes them durable. They survive because the tool that would flag them is the same tool that recommended them.

• Accepting auto-applied recommendations without review. Google Ads can apply certain recommendations automatically, and the defaults skew towards broadening reach — adding keywords, expanding match types, raising budgets. Open the recommendations settings, review what is enabled, and switch off automatic application of anything that changes targeting or budget.
• Judging the account on cost per click. Cheap clicks that do not convert are more expensive than costly clicks that do. Cost per acquisition, and where possible cost per qualified opportunity, are the only figures that connect the account to the business. An account can halve its cost per click and double its cost per customer at the same time.
• Treating negative keywords as a one-off task. Query patterns shift constantly as products, seasons and public conversations change. A negative list built once and never revisited stops earning within a quarter. Thirty minutes a week keeps it current.
• Pausing keywords instead of fixing structure. Pausing a keyword because it converts poorly removes the symptom and the demand together. Frequently the keyword is fine and the ad group is too broad, or the landing page is generic, and the correct fix is a split rather than a pause.
• Sending every ad to the homepage. The homepage is designed to serve everyone, which means it makes a specific promise to nobody. It is simultaneously the most common landing page choice and the largest single drag on both conversion rate and Quality Score in SME accounts.
• Counting every form submission as a conversion. Support requests, supplier enquiries, job applications and existing-customer contacts all arrive through the same form on most SME sites. If they all count as conversions, bidding optimises towards traffic that generates the most form fills rather than the most customers.
• Changing several things at once. Bid strategy, budget, structure and creative changed in the same week produce a result nobody can attribute. Space significant changes at least two weeks apart, and keep a dated log.
• Reacting to a bad week. Most SME accounts do not generate enough weekly conversions for a week’s data to be statistically meaningful. Reacting to normal variance produces a permanent learning phase, which is itself a substantial and entirely self-inflicted source of waste.

Real-world example — a Leeds commercial contractor

A 42-person commercial flooring contractor in Leeds had run the same Google Ads account for four years at approximately £2,400 per month, with enquiry volume gradually declining and cost per enquiry rising from around £60 to just over £140. The instinctive reading was market competition. The evidence said otherwise. The 90-day search terms export showed £1,970 — roughly 27% of the quarter’s spend — going to queries containing “jobs”, “courses”, “flooring apprenticeship”, “how to lay” and “cheapest”. None of those could ever have produced a commercial contract, and none had ever been negated.

Measurement was the larger problem. Three conversion actions were configured, all set as primary: a contact form submission, a phone-number page view, and a brochure download. Two of the three were counted with the “every” method on a lead-generation account. The result was an apparent conversion rate near 9% against a real enquiry rate closer to 3%, and a bidding system that had spent two years learning to buy visitors who browsed the contact page rather than visitors who requested a quotation.

The remediation followed the 90-day sequence: measurement repair first, with the brochure download and phone-page view demoted to secondary and the count method corrected; then a 280-term account-level negative list; then location targeting switched from presence-or-interest to presence-only, which alone removed spend from three regions the business does not serve. Campaign structure was rebuilt in week six, splitting commercial contracts from domestic enquiries so the higher-value work stopped competing for the same budget as the lower-value work. Budget was held flat at £2,400 throughout — the objective was never to spend less.

We assumed the market had got more expensive. It had not. We were paying, every month, for a couple of hundred people looking for flooring jobs and flooring courses, and our reporting was telling us we had twice as many enquiries as we really did. Nobody was doing anything wrong — nobody was looking, and the account never told us to look.

By the end of the quarter, cost per genuine enquiry had returned to the mid-£70s on unchanged spend, with the reported conversion count roughly halving because it was finally counting the right things. That second point is the one worth carrying away: a successful waste-reduction programme very often makes the dashboard look worse before it makes the business better, because the first honest report is the first one that removes the phantom conversions. Any organisation running this exercise should agree that expectation with whoever reads the monthly report, before the report lands.

At-a-glance summary

The key figures, settings and decisions from this guide in one place.

How Cloudswitched works on wasted PPC spend

Cloudswitched approaches Google Ads the way we approach the rest of a client’s technology estate: evidence first, then a sequenced programme with a change log behind it. That means starting with the exports and the conversion action inventory rather than with ad copy, agreeing a frozen baseline so improvement is measurable rather than asserted, and treating measurement repair as a prerequisite to optimisation rather than a task for later. Because we also build and maintain the websites and the underlying infrastructure for many of our clients, the landing page work, the tag configuration and the consent setup sit with the same team — which removes the most common cause of tracking breaking silently after a site release.

Frequently Asked Questions

How much of my Google Ads budget is probably being wasted?

For UK SME accounts spending between £1,000 and £10,000 a month, a first-pass audit typically identifies between 21% and 34% of search budget going to clicks with no realistic path to a sale. Accounts that have run on default settings since launch, or that changed hands between agencies without a proper handover, frequently sit above that range. The only way to know your own figure is to export 90 days of search terms data with cost and conversions, sort by cost descending with conversions set to zero, and read the top rows. That single export usually accounts for the majority of recoverable spend and takes about ten minutes to produce.

What is the fastest change I can make to reduce Google Ads wasted spend?

Switching location targeting from the default presence-or-interest setting to presence-only. It takes two clicks per campaign, takes effect immediately, and in accounts serving a defined geographic area it commonly removes around a tenth of spend without touching a single keyword. The second fastest is applying an account-level negative keyword list covering employment, education and free-intent terms. Neither requires a learning period, neither risks losing genuine demand, and both can be completed in an afternoon. Structural work such as campaign restructuring and bid strategy changes delivers more in total but needs weeks rather than hours.

How many negative keywords should a UK SME account have?

Most well-maintained SME accounts carry between 150 and 400 terms on the account-level list, plus campaign-level and ad-group-level negatives for separation and margin protection. The number matters far less than the structure: three layers, each doing a distinct job, reviewed on a weekly cadence. One caution on match types — negative keywords do not match close variants, so plurals, singulars and common misspellings must each be added explicitly. Negating “job” does not negate “jobs”. Always re-check the search terms report a week after adding negatives to confirm the block actually took effect.

Does improving Quality Score really reduce cost per click?

Yes, because Quality Score feeds Ad Rank, which determines both your position and what you pay for it — two advertisers can hold the same position at materially different prices. Moving high-spend keywords from a score of 4 to 7 typically reduces cost per click somewhere in the region of 15% to 30% at equivalent positions, though the exact effect depends on the auction and the competing advertisers, so treat any precise multiplier with caution. The useful move is diagnostic rather than numerical: add the three component columns to your keyword view, sort by cost, and fix whichever component reads “below average” on your top twenty spenders.

Should I use broad match and Smart Bidding, or stick to phrase and exact match?

It depends almost entirely on conversion volume and signal quality. Smart Bidding needs roughly 30 to 50 conversions per campaign per month to model reliably; below that it is making decisions on thin data. Most UK SMEs sit under that threshold, which is why a hybrid usually works best: run known-intent, revenue-carrying terms in tightly matched campaigns, and run a separate capped discovery campaign on broad match whose only job is to surface new search terms for you to promote or negate. Keep the discovery budget at 10% to 15% of total spend and review it weekly.

Why did my conversions drop after I fixed my tracking?

Because the previous number was wrong. The most common tracking defect is a conversion action firing on a page any visitor can reach, or a lead-generation form counted with the “every” method rather than “one”. Both inflate the count. When you demote soft actions to secondary and correct the count method, the reported figure falls to something closer to reality — often by half. This is a successful outcome that reads as a failure on a dashboard, so agree the expectation with whoever reviews the monthly report before the corrected report lands.

How do I know whether my conversion tracking is broken right now?

Test it as a real visitor rather than in a debugging tool. Open the site in a fresh browser session, interact with the consent banner the way a typical user would, submit the form, and confirm the conversion is recorded in the account. Repeat on a mobile device. Then reconcile the last 30 days of reported conversions against your CRM records for the same period — a variance beyond roughly 10% means one of the two systems is wrong. Re-run this check after every website release, because a rebuild removing a tag is the single most common cause of silent tracking failure.

Is Performance Max wasting my budget?

Not inherently, but it is harder to audit than a search campaign because it spans Search, Shopping, Display, YouTube, Discover and Gmail inventory under one budget, and its reporting does not decompose cleanly into those channels. The specific risk for SMEs is that it harvests brand searches which would have converted anyway, producing a flattering cost per acquisition that overstates its incremental contribution. If you run it, apply account-level negative keyword lists, use brand exclusions deliberately, and read the asset group and search-category reports rather than trusting the headline figure.

How often should I review the search terms report?

Weekly, for about thirty minutes, as a fixed calendar commitment rather than when someone remembers. Query patterns shift continuously as products, seasons and public conversations change, so a negative list built once and left alone stops earning within a quarter. Alongside that, review location and placement reports monthly, and revisit campaign structure and landing page alignment quarterly. Waste is not a defect you repair once; it accumulates, and the cadence is what keeps it from rebuilding to its original level within a year.

Should I just reduce my Google Ads budget instead?

Usually not, if the objective is profitable growth. Cutting spend reduces waste and volume together, in roughly the same proportion, so cost per acquisition stays where it was and you simply buy fewer customers. Cutting waste keeps the volume and removes the cost, which is a different outcome entirely. The exception is an account below roughly 40 on the readiness score in this guide, where measurement is unreliable enough that you cannot tell profitable spend from unprofitable spend — in that situation, holding budget flat while you repair tracking for a fortnight is more sensible than either increasing or decreasing it.

Can I do this myself or do I need an agency?

The first pass is genuinely doable in-house. The twelve-point checklist in this guide needs three to four hours on a single-country account with fewer than ten campaigns, and every step uses data Google already gives you. Where external help earns its fee is the harder half: campaign restructuring without losing historical performance, offline conversion import from a CRM, margin-aware bidding, and landing page work. A reasonable pattern is to run the diagnostic yourself so you know what the account actually looks like, then buy specialist time for the specific problems the diagnostic found.

What should I ask an agency that manages my Google Ads?

Four questions cover most of it. When did you last review the search terms report, and what did you negate? How many primary conversion actions are configured, and what does each one represent? Is location targeting set to presence-only? And what is our cost per acquisition on non-brand traffic specifically, separated from brand? A capable manager answers all four immediately with numbers. Vague answers, or reporting that blends brand and non-brand into a single averaged figure, are the most reliable warning signs available to a non-specialist.

Related reading

• Local SEO for UK Businesses: A Practical Guide to Ranking in Google’s Local Pack in 2026
• Core Web Vitals and Conversion: A UK Business Guide to Faster, More Profitable Websites in 2026
• When Does a UK SME Need a Virtual CIO? A Practical Guide to IT Strategy Without a Full-Time Hire in 2026
• In-House vs Outsourced IT Support: A UK SME’s Decision Framework for 2026
• Network Segmentation for UK SMEs: A Practical Guide to Isolating Risk Without Breaking Productivity in 2026