Quality Score & Negative Keywords: Two Levers Most Accounts Ignore

Paid Media

30 Sept

Most accounts get reviewed by adjusting bids and rewriting headlines. Meanwhile two settings sit untouched: a diagnostic that tells you exactly which part of the funnel is weak, and a list that stops you paying for searches you never wanted. Both are free to change. Both move cost more than a bid tweak.

What Quality Score is & what it is not

Google Ads Help defines Quality Score as "a diagnostic tool meant to give you a sense of how well your ad quality compares to other advertisers". It is reported at keyword level on a scale of 1 to 10, and each of its three components is graded as above average, average or below average relative to other advertisers over the previous 90 days.

Then Google says the thing most advertisers have never read. "Quality Score isn't an input in the ad auction." It also states that Quality Score "isn't a key performance indicator and shouldn't be optimised or aggregated with the rest of your data".

That single sentence rules out a large amount of common agency practice. The number in your keyword table is not what the auction reads. It is a summary of how your ad quality compares, offered so you can find the weak component and fix it. Ad quality does matter in the auction, as we will come to. The reported score is a mirror, not the machine.

The three components

Google Ads Help lists three, and each points at a different fix.

Expected click through rate is "the likelihood that your ad will be clicked when shown". Below average here usually means the ad is not answering the search. The fix lives in the ad, not the bid.

Ad relevance is "how closely your ad matches the intent behind a user's search". Below average here is most often a grouping problem. One ad group carrying twenty loosely related keywords cannot write one ad that closely matches all of them. Splitting by intent fixes more relevance problems than rewriting copy does.

Landing page experience is "how relevant and useful your landing page is to people who click your ad". Below average here is rarely a page speed issue in isolation. It is usually that the page is a general service page and the search was specific, so the visitor has to hunt for the thing they searched for.

Read the three together and they read like a diagnosis. Below average on relevance and landing page, average on click through rate, tells you the ad is writing cheques the page does not honour.

Where Ad Rank actually decides the outcome

Ad Rank is the mechanism, and Google Ads Help sets out six factors: your bid, ad and landing page quality measured at auction time, Ad Rank thresholds, the competitiveness of the auction, the context of the search, and "the expected impact from your ad assets and other ad formats".

Two details in that documentation change how you should think about the account.

First, timing. Google states that "Ad Rank is calculated every time a user does a search and is recalculated for different positions on the search results page". Quality is assessed live, per auction. A keyword level score calculated over 90 days cannot be the input to something recalculated per query, which is exactly why Google says it is not.

Second, cost. Google states plainly that "higher quality ads can often lead to lower CPCs. That means that you pay less per click when your ads are higher quality." Google also describes Ad Rank thresholds as minimum standards an ad must meet to show at all, with higher thresholds for top of page positions. Quality is not only a discount. It is an eligibility gate.

Three misreadings that cost money

Averaging Quality Score across the account. Google explicitly says it should not be aggregated. An account average of 6.4 is a number with no referent. It cannot be acted on and it hides the twelve keywords that are genuinely below average on landing page experience.

Chasing tens on keywords with no volume. The components are relative comparisons over 90 days. A keyword with a handful of impressions produces an unstable comparison. Spend your attention on the keywords carrying real cost.

Treating a low score as a bid problem. Raising a bid on a keyword whose landing page experience is below average buys more expensive clicks to the same weak page. The score is telling you which of three things to repair. Repair that thing.

The search terms report is the real input

Quality Score tells you how well you are answering the searches you already match. The search terms report tells you which searches you are matching in the first place, and this is where most wasted spend lives.

Google Ads Help describes the report as showing "actual searches within the Search Network" that triggered your ads, and draws the distinction that matters: "A search term is a word or set of words that a person enters when searching on Google" while a keyword is what you added to an ad group. They are not the same list and in a broad match account they can diverge sharply.

Google's own match type documentation explains why. Broad match "may show on searches that are related to your keyword, which can include searches that don't contain the direct meaning", and Google says the system considers the user's recent search activity, your landing page and asset content, and the other keywords in the ad group. Phrase match shows on searches that include the meaning of your keyword. Exact match shows on searches with the same meaning or intent.

One caution before you start reading. Google states that "some search terms that don't have enough query activity are omitted from the search terms report in order to keep with our standards on data privacy". The report is not a complete census of every query, so the cost you can see is a floor, not a ceiling.

Negative match types do not work like positives

This is the part accounts get wrong most often. Google Ads Help sets out three behaviours.

Negative broad match, the default, blocks the ad when the search contains all of your negative terms in any order. If the search contains only some of them, the ad can still show.

Negative phrase match blocks the ad when the search contains your exact terms in the same order, even if other words are present.

Negative exact match blocks the ad only when the search is exactly your terms with no additional words.

Then the line that catches everyone: "Negative keywords won't match to close variants or other expansions." Positive keywords expand. Negatives do not. If you block "free" you have not blocked "freebie". If you block "job" you have not blocked "jobs". Plurals, misspellings and synonyms have to be added by hand.

List hierarchy & a weekly workflow

Three tiers, each with a job.

Account level negatives, which Google states "automatically apply to all eligible search and shopping inventory in relevant campaign types", are for terms that are wrong for the whole business. Recruitment terms, "free", "DIY", "how to make", the names of unrelated products with the same word in them.

Shared negative keyword lists are for themes that apply to several campaigns but not all. Google documents a limit of 5,000 negative keywords per list and up to 20 lists per account, with a separate account level limit of 1,000 negative keywords. Updating one list updates every campaign it is applied to.

Campaign and ad group negatives are for steering, not blocking. This is where you stop a brand campaign eating generic traffic, or stop one service ad group cannibalising another.

The recurring routine matters more than any single clean-up. Google's guidance for pulling negatives from the report is to go to Campaigns, then Insights and reports, then Search terms, select the terms to exclude, and add them as negatives to an ad group, campaign, an existing list or a new one.

Filter the search terms report to the last seven days, sorted by cost, with conversions set to zero.
Read the top fifty terms only. Beyond that the cost per term is usually too small to justify the time.
Sort each one into wrong intent, wrong product or wrong stage. Wrong intent goes to the account level or a shared list. Wrong product and wrong stage go to the campaign.
Choose the negative match type deliberately. Exact for a single term you want blocked in that precise form, phrase for a recurring word cluster, broad only when you are confident every part of the phrase is disqualifying.
Add the plurals, misspellings and synonyms yourself, because Google will not.
Check last week's additions did nothing you did not intend. Look for keywords whose impressions fell to zero.
Log what you added and the date. In six months someone will ask why a term stopped serving.

Twenty minutes a week, done every week, beats a four hour audit once a quarter.

What the wasted spend is worth, in pounds

Take an account spending £4,000 a month on search at an average cost per click of £2.20. That is roughly 1,818 clicks.

Run the workflow above and, in a typical account that has never had a negatives routine, somewhere between 10 and 20 per cent of clicks turn out to come from search terms that could never have converted: job seekers, students, people looking for free versions, people searching for a similar product you do not sell. Take the middle of that range, 14 per cent. That is 254 clicks, and at £2.20 each it is £559 a month, or £6,708 a year.

That money does not disappear when you block it. It redeploys into the terms that already convert. At an account cost per acquisition of £48, £559 a month buys roughly 11 additional enquiries every month, from the same budget, with no change to bids, creative or landing pages.

Then apply the Quality Score reading to what remains. If the diagnostic points at landing page experience on your three highest cost keywords, and repairing those pages produces the lower cost per click Google describes as a typical outcome of higher quality ads, the two levers compound. Neither one required an increase in budget.

What to do next

Export the last 90 days of search terms, sorted by cost with zero conversions. Read the top fifty. Total the cost. That number is your annual saving divided by four.
Create one account level negative list this week for terms that are wrong for the entire business, and one shared list per service line.
Go through your existing negatives and add the plurals, misspellings and obvious synonyms. Assume nothing has been expanded for you.
Filter your keyword table to keywords with real cost and a below average component. Fix the component the score names, not the bid.
Put a recurring twenty minute slot in one person's calendar, same day every week, and give it an owner.

Neither of these levers looks impressive in a monthly report, which is precisely why they stay untouched in most accounts. They also cost nothing but attention. If you would rather have someone run that routine properly every week, it sits inside our paid media work.

Sources

google adsquality scorenegative keywordssearch terms reportppc

wevisualise designs https://wevisualise.co.uk