How Many Keywords Should Be In One Ad Group? Fewer Than You Have.

Google lets you put 20 keywords in an ad group.

Most people read that as a recommendation. It’s a technical limit, in the same way a lift has a maximum weight rather than a target weight.

The recommended number is between 3 and 10, sharing a single search intent.

We regularly take over accounts with forty, sixty, sometimes over a hundred keywords sitting in one ad group. Nothing is broken. Everything still runs.

It’s just costing them up to four times more per click than it needs to.

This post explains why that happens, how to tell if it’s happening to you, and how to repair an account that’s already grown wild.

What Keyword Bloat Actually Is

Keyword bloat is what happens when an ad group accumulates keywords over time, and nobody ever removes any.

It’s rarely a decision. It’s usually a series of small, sensible-looking additions.

You launch with eight keywords. You spot a good search term and add it. A month later, you add three more. Somebody suggests a variation, so in it goes.

Two years later that ad group has sixty keywords covering nine different things people want, and one set of ads trying to speak to all of them.

Nothing about that process feels wrong at any single step. That’s what makes it so common.

And here’s the important bit. Bloat doesn’t announce itself. Your ads still show. Your budget still spends. The account looks busy and healthy.

The cost is hidden in what you pay per click.

The Mechanism: One Ad Cannot Serve Forty Intents

To see why this matters, you have to understand where ads actually live.

Ads sit at ad group level, not keyword level. Every keyword in an ad group is served by the same ads.

So if your ad group holds “emergency plumber”, “bathroom installation”, “boiler service” and “underfloor heating”, one set of ads has to work for all four.

Try writing that headline. You end up somewhere like “Trusted Local Plumbers, Call Today”.

It’s not wrong. It’s just not about anything.

Now compare it to an ad group holding only emergency terms. The headline becomes “Emergency Plumber, 60 Minute Response” and the person with water coming through the ceiling recognises themselves immediately.

That’s the whole mechanism. A responsive search ad covering twenty, thirty or fifty different keyword intents tries to be relevant to everything, and ends up specific to nothing.

Relevance isn’t a scoring abstraction. It’s whether the searcher sees their own problem in your headline.

What Bloat Actually Costs You

Now the money, because this is where it stops being a matter of tidiness.

Google gives every keyword a Quality Score out of ten, driven largely by expected click-through rate, ad relevance and landing page experience. We covered how that score works in detail in an earlier post, so here we’ll stick to the consequence.

The consequence is your cost per click.

A keyword sitting at Quality Score 1 to 3 can cost **up to 400 per cent more per click** than a keyword at the score 5 baseline.

A keyword at Quality Score 10 can attract a discount of **up to 50 per cent**.

Same click. Same ad position. Wildly different price.

Put that on a real budget. Say you spend £800 a month at an average £4 a click. That’s 200 clicks.

Lift your average Quality Score enough to move your cost per click to £3, and the same £800 buys 267 clicks. Sixty-seven extra clicks a month, no extra spend, purely from structure.

That’s the argument. Not that bloated accounts are untidy, but that they pay a premium on every single click for the privilege.

Three Kinds Of Bloat, Three Different Fixes

Most people think bloat means one thing. There are actually three, and most accounts have at least two.

Type one: too many keywords in an ad group. The classic. Nine intents, one set of ads, nothing specific to anyone. Fixed by splitting the ad group.

Type two: too many ad groups, each too thin. The opposite failure, and usually the result of somebody reading about single-keyword ad groups. Forty ad groups, two keywords each, three conversions a month spread across all of them. Fixed by merging.

Type three: the same keyword in several places. Duplicates scattered across campaigns, usually because nobody kept a master list. Fixed by choosing one owner.

Type three deserves a note, because there’s a myth attached to it. Duplicate keywords don’t bid against each other and inflate your prices. Google picks one and runs it.

The real damage is quieter. Your performance history fragments across several copies, so no single instance accumulates enough data to build a strong Quality Score. Your reporting also becomes unreliable, because the same search term shows up in three places.

Pick one home per keyword and delete the rest.

Single Keyword Ad Groups Are Over

This needs saying plainly, because SKAGs are still being recommended by people who stopped paying attention around 2021.

The idea was sound in its time. One keyword per ad group, so the ad could match that keyword perfectly. Maximum relevance, maximum Quality Score.

It doesn’t work now, for two reasons.

Close variants killed the isolation. Since exact match started matching on meaning rather than wording, a single keyword serves dozens of different searches anyway. The clean one-to-one relationship SKAGs depended on no longer exists. You’ve built a wall around a keyword that leaks.

Smart Bidding needs concentrated data. Split your conversions across sixty tiny ad groups, and none of them has enough signal for the algorithm to learn from. You’ve starved the system that decides your bids.

So the pendulum has swung. The modern answer is neither one keyword nor fifty. It’s a small group of keywords that mean the same thing.

Three to ten, one intent, one set of ads that speaks to it directly.

The Test That Settles Any Ad Group

Here’s the only diagnostic you need, and it takes about thirty seconds per ad group.

Try to write one headline that would genuinely appeal to somebody searching for any keyword in that group.

If you can do it without going vague, the group is fine.

If you find yourself reaching for something like “Quality Service, Great Prices”, the group holds more than one intent and needs splitting.

That’s it. No tool, no score, no spreadsheet.

The test works because it’s the same job your ad has to do every time it serves. If you can’t write the headline, your ad can’t either.

A related version for the landing page. Could every keyword in this ad group sensibly land on the same page? If two keywords need different pages, they need different ad groups.

How To Audit Your Own Account

Twenty minutes, and you’ll know exactly where you stand.

Open the Ad groups tab and add the keyword count column if it isn’t showing. Sort by keyword count, highest first.

Anything over 15 keywords goes on the split list. Anything over 30 goes at the top of it.

Now sort the other way. Anything with 1 or 2 keywords goes on the merge list.

Next, sort your ad groups by cost. Work top down, because the expensive ad groups are where a fix pays back fastest. A perfectly structured ad group spending £4 a month is not worth an afternoon.

Then check ad counts. Every ad group should hold at least two ads so they can be tested against each other. Ad groups with one ad go on a third list.

Finally, export your keyword list to a spreadsheet and sort alphabetically. Duplicates jump straight out.

Three lists and a duplicate check. That’s your repair plan.

How To Split A Bloated Ad Group

Don’t start deleting. Start grouping.

Export the keywords from the bloated ad group into a spreadsheet.

Now sort them by what the searcher actually wants, not by wording. This is the part that takes judgement, and it’s the part that matters.

A plumber’s sixty-keyword ad group usually splits along lines like these: emergency and urgent work. Boiler repair and servicing. Bathroom installation. Leaks and drainage. Heating systems.

Five clean intents. Five ad groups.

Once you have your groups, build them properly rather than halfway.

Each new ad group gets its own name describing the intent, its own 3 to 10 keywords, at least two ads written specifically for that intent, and a landing page that answers that specific need.

Then move the keywords across rather than recreating them, so you keep the performance history you’ve already paid to build.

Pause the old ad group. Don’t delete it for a month, in case you need to look at what was in it.

One warning about pace. Split one ad group at a time and leave a fortnight between them. Restructure everything in one afternoon, and you’ll never know which change did what, and you’ll push several campaigns back into a learning period at once.

A Worked Example: Sixty Keywords Into Five Ad Groups

Here’s the job done properly, on a typical local trade account.

The starting position. One campaign, one ad group called “Plumbing”, sixty-two keywords, two ads. Spending £700 a month. Average cost per click £4.20, so roughly 167 clicks. Around eight enquiries.

The two ads read “Trusted Local Plumbers” and “Plumbing Services, Call Today”. Both entirely reasonable. Neither about anything.

Export the keywords and sort by what the searcher wants.

Emergency, 14 keywords. Emergency plumber, burst pipe, no hot water, leak repair urgent. Somebody with a problem happening right now.

Boilers, 17 keywords. Boiler repair, boiler service, boiler replacement, boiler not working. A specific appliance, often a considered purchase.

Bathrooms, 12 keywords. Bathroom fitting, bathroom installation, shower installation, new bathroom. Planned work, longer decision, higher value.

Drainage, 9 keywords. Blocked drain, drain unblocking, blocked toilet, drain cleaning. Urgent but different from a water leak.

Heating, 10 keywords. Radiator installation, underfloor heating, central heating repair, power flush.

Sixty-two keywords, five intents. Nobody planned it that way. It just accumulated.

Now the rebuild. Five ad groups, each with its own name, its own keywords, at least two ads written for that specific intent, and its own landing page.

The emergency ad group gets headlines like “Emergency Plumber, 60 Minute Response” and “Burst Pipe? We’re On Our Way”. Somebody watching water come through a ceiling recognises themselves instantly.

The bathrooms ad group gets something completely different. “New Bathroom Fitted In 5 Days” and “Free Bathroom Design Visit”. A different person, a different worry, a different promise.

Neither of those headlines could have existed in the old structure, because neither would have suited the other four intents.

Look at what changes. Same sixty-two keywords. Same £700 budget. Same business, same vans, same team.

The only difference is that each searcher now sees an ad about their own problem, and lands on a page that answers it.

That’s what the Quality Score discount is actually paying for. Not tidiness. Recognition.

Where The Keywords You Cut Should Go

Splitting is only half the job. Some keywords shouldn’t be moved anywhere, and knowing the difference matters.

Work through the leftovers in four piles.

Keep and move. The keyword converts, or has genuine intent and enough volume. It goes into whichever new ad group matches its meaning.

Demote to phrase or exact. The keyword is fine but broad, and it’s been pulling in loose traffic. Tighten the match type rather than removing it. Add the tighter version alongside and pause the loose one, so the history survives.

Turn into a negative. Some of what’s sitting in your keyword list should never have been there. If a keyword has spent money over 90 days with no conversions and the searches behind it were never going to buy, it belongs on your negative list instead.

Park. Low volume, no data, but plausible. Pause it rather than deleting. You lose nothing, and you can revive it if the market changes.

The rule for the whole exercise: pause before you delete, and check the search terms report before you decide anything.

We regularly find a keyword that looks dead at ad group level but is quietly converting on one specific search term nobody noticed. Delete it, and you lose that search along with it.

Above The Ad Group: When To Split A Campaign

Ad group structure is where most of the relevance lives, but it sits inside campaign structure, and campaign structure controls the money.

Two things live at campaign level and nowhere else. Budget and location targeting.

That gives you the rule. Split into a separate campaign when you need to control spend or geography separately. Split into a separate ad group when you need to control the message.

So the plumber above keeps all five ad groups in one campaign if they’re happy for budget to flow between them.

They split emergency work into its own campaign the moment they want to guarantee it a budget, because emergency work is their best margin and they can’t have bathroom searches eating the money on a busy Tuesday.

Same logic for geography. If you’ll travel forty miles for a bathroom fit but only ten for a leak, those are different campaigns, because the radius is a campaign setting.

How many ad groups per campaign? There’s no hard limit worth quoting. In practice, when a campaign holds more than about ten ad groups, budget starts spreading too thin for any of them to gather useful data. That’s the signal to split.

One caution about over-splitting at campaign level. Every campaign has its own budget, and small budgets learn slowly. Ten campaigns at £3 a day each is a worse account than one campaign at £30, almost every time.

The Bloat You Can’t See

There’s a version of this problem that doesn’t show up in your keyword count at all, and it’s worth checking while you’re in there.

Your keyword list might be beautifully tidy. Three to ten per ad group, clean intents, well-written ads.

And your search terms report might still show that one ad group is serving forty different things.

That’s close variants doing the work your structure was supposed to do. Google has decided that a range of searches all mean roughly what your keyword means, and it’s serving them all through the same ads.

So run the check from the other end. Pick your best ad group. Open the search terms report, filter to that ad group, and read what it actually served over 90 days.

If those searches split into two or three clearly different needs, your ad group is bloated in practice even though it looks tight on paper.

The fix is the same. Add the distinct intents as their own keywords in their own ad groups, then let the tighter keywords pick up that traffic instead.

This is the check almost nobody runs, and it’s why some accounts with textbook structure still have poor ad relevance.

What To Expect After You Split

The pattern is consistent enough to predict.

Weeks one and two. Performance dips. You’ve reset the learning on those ad groups, and the new ones have no history. This is normal, and it’s the reason people abandon restructures halfway.

Weeks three and four. Click-through rate starts climbing because your ads finally match the searches.

Weeks five to eight. Quality Score follows click-through rate up, and cost per click starts drifting down.

Weeks eight to twelve. The full effect lands. More clicks for the same budget, and better matched clicks at that.

Quality Score is a lagging indicator. It’s built on accumulated data, so it can’t move the day after you make a change.

Give any restructure a full quarter before judging it. This is the single most common reason a good restructure gets reversed at week three.

The Mistakes We See Most

Treating Google’s 20 keyword limit as the target. It’s a ceiling, not a recommendation.
Splitting into single keyword ad groups and starving Smart Bidding of data.
Restructuring the whole account in one weekend, then having no idea what worked.
Splitting the ad group but leaving the same generic ad in all the new groups. The split does nothing if the ads don’t change.
Sending every new ad group to the same landing page. Half the relevance benefit lives on the page.
Deleting keywords instead of moving them, and binning the performance history with them.
Chasing Quality Score as a number rather than as a symptom. Fix relevance, and the score follows.
Judging a restructure at three weeks, when the metric you’re watching is a lagging one.

Your Twenty Minute Job This Week

One sitting, no tools.

Minutes one to five. Ad groups tab, add the keyword count column, sort highest first. Write down everything over 15.

Minutes six to ten. Sort by cost. Find the overlap between expensive and bloated. That’s your first job.

Minutes eleven to fifteen. Take your single worst offender and try to write one headline that suits every keyword in it. When you can’t, you’ve proved the case.

Minutes sixteen to twenty. Export those keywords and group them by intent in a spreadsheet. You now have your new ad groups on paper.

Build them next week. One at a time.

How Bloat And Match Types Compound

One last connection worth making, because these two problems feed each other and most accounts have both.

Loose match types widen what each keyword serves. Bloated ad groups widen what each ad has to cover. Put them together and the damage multiplies rather than adds.

Picture the worst case, which is genuinely common. Forty broad match keywords in one ad group, with one generic ad.

Each of those forty keywords is pulling in a wide circle of loosely related searches. All of that traffic lands on the same two ads. Those ads have to speak to forty intents already, and now they’re speaking to several hundred actual searches.

Click-through rate falls through the floor. Quality Score follows it. You’re paying the premium on every click, on traffic that was already loosely matched.

Now the reverse, which is the account you want. Tight match types, tight ad groups. Each keyword serves a narrow set of searches, and each ad group’s ads speak directly to those searches.

Click-through rate rises. Quality Score rises. Cost per click falls. And because the traffic was better matched to begin with, more of those cheaper clicks turn into enquiries.

The order to fix them matters, and most people get it backwards.

Fix the structure first. Split the ad groups, write the specific ads, sort the landing pages. Then tighten the match types.

Do it that way around, and the tighter keywords have properly relevant ads waiting for them from day one. Do it the other way, and you tighten your traffic into the same vague ads, which wastes half the benefit.

Structure, then match types, then negatives. That’s the sequence.

Frequently Asked Questions

How many keywords should be in one ad group?

Between 3 and 10 keywords sharing a single search intent. Google permits up to 20, but that’s a technical limit rather than advice. The real test is whether one ad can speak convincingly to every keyword in the group. If it can’t, you have more than one ad group.

Does having too many keywords hurt Quality Score?

Indirectly, and a lot. Quality Score is driven largely by expected click-through rate and ad relevance. One ad serving many intents can’t be specific to any of them, so click-through rate falls, and the score follows. The keyword count isn’t penalised. The vagueness it forces on your ads is.

How much does a low Quality Score cost?

A keyword at 1 to 3 can cost up to 400 per cent more per click than one at the 5 baseline. A keyword at 10 can attract up to a 50 per cent discount. Same click, same position, very different price, which is why structure is a budget question rather than a tidiness question.

Are single keyword ad groups still worth using?

Generally not in 2026. Close variants mean one keyword already serves dozens of searches, so the isolation SKAGs were designed to create no longer exists. They also spread your conversion data thin across many ad groups, which starves Smart Bidding of the signal it needs.

How do I know if my ad group has too many keywords?

Write one headline that would genuinely appeal to anyone searching for any keyword in the group. If you can do it and it stays specific, you’re fine. If you end up at something like “Quality Service, Great Prices”, the group holds several intents and should be split.

Do duplicate keywords across campaigns cause problems?

They don’t bid against each other, because Google picks one. The damage is that performance history fragments across copies, so no single instance builds a strong Quality Score, and your reporting stops being trustworthy. Give every keyword one home.

Should I delete my worst keywords or just pause them?

Pause first, and review after a month. Deleting removes the history, and occasionally a keyword that looks poor at ad group level is actually converting on a search term you hadn’t noticed. Check the search terms report before removing anything.

How long before a restructure shows results?

Expect a dip for a fortnight, click-through rate improving from week three, and cost per click responding from around week five. The full effect usually lands between weeks eight and twelve. Quality Score is a lagging indicator built on accumulated data, so patience is part of the method.

Fewer Keywords, Better Ads, Cheaper Clicks

Keyword bloat is one of the few problems in Google Ads that costs you money every single day and never once shows up as a failure.

Your ads run. Your budget spends. The graphs look normal. You simply pay a premium on every click because your ads can’t be specific enough to earn a better price.

The fix isn’t clever, and it isn’t expensive. It’s a spreadsheet, some honest grouping by intent, and the patience to do one ad group at a time.

Start with the test. Open your biggest ad group and try to write one headline that suits every keyword in it.

If you can’t, you already know what to do next.

And be fair to yourself about how it got this way. Nothing in Google Ads warns you when an ad group gets too big. There’s no prompt saying these keywords want different ads.

You added a keyword because you spotted a good search, which was the right instinct. You did it fifty times over two years and ended up somewhere nobody chose.

That’s the tool’s design, not your carelessness.

If you’d rather someone went through the account and did it properly, that’s what we do at SF Digital.

Did You Enjoy This Blog Post?

I hope you enjoyed this blog post, and thank you so much for being here. We also upload videos to our YouTube channel every weekday. Please subscribe so you are one of the first to be notified.

If you enjoyed this blog, you may also like:

SF Digital Studios

Search This Blog