LOW KEYWORD QUALITY GOOGLE ADS

Every account I inherit has one: a keyword that drives a meaningful share of the spend, sits at a 3 or a 4, and has been that way for a year. Everyone knows about it. Nobody has fixed it. The generic advice write better ads, improve the landing page, add negative keywords has already been applied twice and the number hasn't moved.

What follows is what I actually found when I stopped treating low keyword quality google ads as a copywriting problem and started treating it as a structural one. Four causes, in the order I keep running into them, and what each one required.

Reading the google ads keyword quality score properly

The composite number is almost useless on its own. Google's official documentation on Quality Score is explicit that it's a diagnostic tool and not an auction input, and that it breaks into three components: expected CTR, ad relevance, and landing page experience, each rated above average, average, or below average.

Two things follow from that, and both change how you work. First: the score is built on historical impressions for exact searches of the keyword, which is why switching match types does nothing to it. I've watched people move a term from broad to phrase expecting a lift and get an identical number three weeks later.

Second: a 4 with below-average ad relevance and a 4 with below-average landing page experience are unrelated problems wearing the same label. So the first thing I do is add the three component columns, plus their historical versions, and ignore the headline number entirely. The composite tells you there's a problem. The split tells you which one.

The four causes I actually found

1. The keyword was in the wrong ad group

By far the most common, and the one nobody wants to hear because it means restructuring. A term gets added to whichever group looked closest at the time. Six months later the ads in that group speak to a slightly different intent, and expected CTR reads below average because users see the ad and don't click it.

The tell: the keyword's score is materially worse than the other keywords in the same group. If the whole group is at 4, that's an ad problem. If one term is at 3 while its neighbours sit at 7, that term doesn't belong there. This is why my keyword targeting decisions now start from intent grouping rather than semantic similarity terms that look alike often convert differently.

2. Expected CTR was poisoned by old history

The component is built on past performance for that exact search. If the keyword ran for eight months under a headline that didn't match it, that history is still weighing on the score after you fix the ad. The number lags the fix by weeks, sometimes longer on low-volume terms.

I've seen accounts where the ads were rewritten, the score stayed flat for a month, and the team concluded the rewrite failed. It hadn't. It just needed enough new impressions to outweigh the old ones. On a term getting forty exact-match impressions a week, that's a long wait.

3. The landing page was relevant to the business, not to the query

Landing page experience is the component where I see the most self-deception. The page is good. It converts. It's professionally built. And it's still below average, because it answers a broader question than the one in the query.

A concrete pattern: a service keyword pointing at a category page listing eight services. The page is genuinely useful and genuinely relevant to all eight. For the specific query, the visitor has to do work to find their answer. Splitting to a dedicated page moved that component from below average to average within about six weeks, and it's the single fix I've seen produce the most reliable change.

4. Nobody had checked whether the score even existed

A dash in the Quality Score column is not a low score. It means Google doesn't have enough exact-match searches to compute one. I've been handed "our quality score problem" spreadsheets where a third of the rows were dashes being read as zeros.

That's not a quality problem, it's a volume problem, and it usually indicates an over-granular keyword list the same pattern that produces a wall of low search volume flags in the keyword status column. Different label, same root cause.

How I raised it on the keyword that mattered

Isolate before you edit anything

If the term carries real spend, I pull it into its own ad group rather than rewriting the group's ads. It's reversible, it doesn't disturb the other keywords, and it makes the result readable you know any change came from this move and not from something else.

Write the ad to the query, not to the offer

Ad relevance is a match between the search intent and the ad. The instinct is to write the most persuasive ad. The correct move is to write the most literal one first, get the component to average, then optimise for persuasion inside that constraint. I've reversed this order and it costs weeks.

Match the page to the query's specificity

Not a better page a narrower one. If the query names a specific service, city, or product, the page should name it in the first screen. This is the fix that survives; ad copy changes get overwritten by the next person, page structure usually doesn't.

Then wait, and read the historical columns

Segment by day and watch the historical component columns rather than the current composite. You'll see a component flip from below average to average well before the headline number changes. That's your confirmation the fix landed. If you're only watching the 1–10, you'll abandon a working change.

Keep bids out of it

Raising bids is the reflex and it's the wrong lever. Under Smart Bidding it isn't even a lever you hold directly the algorithm is setting per-auction bids from conversion signals, so there's no fixed number to raise. Chasing impression share on a low-quality term buys expensive clicks that then perform badly, which is a good way to make things worse. If cost is the actual concern, that belongs in a keyword bidding conversation, not a quality one.

What the pattern of low scores tells you

The distribution across the account is more informative than any single keyword, and it reads like a build history.

Low scores concentrated in two or three ad groups means someone kept adding keywords to existing groups instead of creating new ones. Common on accounts that changed hands, where each manager inherited a structure they didn't want to touch and appended to it. The fix is structural and it's larger than the keyword.

Low scores scattered evenly across the account is a different animal. That's usually a site-wide landing page issue or a fundamental mismatch between what the account bids on and what the business actually sells. No amount of ad rewriting fixes it.

Below-average landing page experience across the board while ad relevance is fine tells you the account was built by someone who owned the ads but not the website. I've seen this a lot in agency handovers the copy is sharp, the destination was never in scope.

One low keyword surrounded by high ones is the best case and the most actionable. It's a placement error, not a quality problem, and it's usually a twenty-minute fix once you accept that's what it is.

The inference that matters: a low google ads keyword quality score is rarely about the keyword. It's about the distance between what someone searched, what you showed them, and where you sent them. Every real fix I've made closed one of those two gaps.

What I stopped doing

Chasing 10s. A 7 on a keyword that converts is finished work. The number is diagnostic, not a KPI, and treating it as a target produces optimisation for its own sake.

Averaging Quality Score across the account. An account-level average is a meaningless figure that hides exactly the concentration pattern you need to see. I've watched teams report it monthly and learn nothing from it.

Rewriting ads more than once before waiting. Two rewrites inside three weeks means you can't attribute the outcome to either. One change, then patience.

Deleting low-scoring keywords to tidy the account. Removing the row removes the evidence. If the term has intent and volume, the low score is information about your structure, not a reason to lose the term.

The practical takeaway

Ignore the composite number and read the three components. Then ask one question: is this keyword scoring worse than its neighbours, or is the whole group low? The first is a placement problem and moves in a day. The second is an ad or page problem and moves in weeks.

Isolate the keyword, write the ad literally before writing it well, narrow the landing page to the query's specificity, and then watch the historical component columns rather than the headline score. In every account where low keyword quality google ads was a persistent problem, the cause was structural and the generic advice had been failing precisely because it was aimed at the wrong layer.