Skip to the tool
Just Better Marketing PPC Tools

Free tool · Google Ads · Runs in your browser

Quality Score diagnostics: what the number means, and what it does not

Quality Score is the most misread metric in paid search. It is treated as a performance grade, so accounts are rebuilt to chase a 10 that was never a promise. It is not. Google describes it as a diagnostic of predicted relevance, and says outright that it does not predict performance. That makes it genuinely useful — as a map of where the ad and the page disagree with the searcher. It just cannot answer the question people actually ask of it.

✔ The documented bands, nothing invented ✔ No predicted cost, conversion or return ✔ Says which keywords are fine as they are

1. Your Quality Score

The number from the keyword status view, and the three diagnostics Google shows beside it.

One decimal place, as Google reports it. Leave it blank if no score is shown.

The diagnostic wording differs between Search and Display/Video, so the tool asks which you have.

2. The three diagnostics

Exactly as Google words them for you. “Not sure” is a valid answer and a common one.

Search campaigns use above or below average. Display and Video use a five-step scale from very low to high, and this tool will explain the wording you actually have.

Headline

Ready
Your band - -
Diagnostics to act on - -
Predicted performance Not predicted -

The per-diagnostic explanation, the ordered fix list and the CSV export unlock below.

Unlock the full diagnostic

Free

Get each of the three factors explained separately, the documented causes for the rating you have, an ordered fix list, what raising your Quality Score would cost you in clicks, and a CSV of the whole diagnostic.

Already unlocked? Reset the gate.

The four bands, and what each one is for

These are the documented edges, and they are the only numbers in this tool that come from Google. Nothing is rescaled or reinterpreted.

  • Above 7 — Excellent. Google reports relevance as above average on the factors it can measure. There is no published ceiling of exactly 10 in the band definitions, so do not treat 10 as the target.
  • 4 to 7 — Good. The working range of most healthy accounts. Most profitable accounts have never sat above 7 and do not need to.
  • 1 to 4 — Needs improvement. Worth investigating, and worth it only where the keyword is supposed to convert.
  • Below 1 — Poor. Rare, and usually a sign of a genuine mismatch between the keyword, the ad and the page, or of an ad that is barely eligible to show.

Why a single number compresses three separate things

Two keywords can share a Quality Score of 5 for entirely different reasons. One may have strong ad relevance and a page that takes nine seconds to load. The other may have a clean, fast page and an ad that does not answer the query. The number is identical and the work is opposite, which is why the diagnostic beside it matters more than the total.

What a Quality Score is made of

Quality Score = a combination of three assessments

1. Ad relevance

2. Expected click-through rate

3. Landing page experience

All three are estimates based on historical signals, and none of them is a measurement of a conversion. The combination is not published as a formula with weights, and this tool will not invent one.

What it is not, precisely

It is not a conversion rate, a profitability measure, a domain authority score, or a statement about your competitors. It is Google's estimate of how relevant your ad and page are to a particular query, which is a statement about one impression at a time.

The mistake this tool exists to prevent

Rebuilding an account to raise a metric that was never a forecast of profit, and discovering the profit fell. The score is worth fixing where it disagrees with what customers do. It is not worth fixing everywhere.

Ad relevance: does the ad answer the query

Short answer

Google is judging whether your ad text is relevant to the keyword, and it does this per keyword, not per ad group. A good ad in the same ad group as a poor one does not raise the poor one's score.

The usual causes of a below-average rating are an ad group that mixes several unrelated themes, a keyword whose intent your ad copy does not address, and a keyword so broad that no single ad could satisfy it.

What to change

  • Tighten the ad group to one theme. A group selling both work boots and wedding photography cannot produce a relevant ad for every keyword in it.
  • Make sure the headline carries the core idea of the query. Not the exact phrase — the searcher's actual subject, in your words.
  • Add ad copy variants that answer different intents in the same group, rather than one ad stretched across both.
  • If the product genuinely cannot satisfy the keyword, pause it. A high score on a keyword that does not convert is a trap, not an achievement.

Expected click-through rate: will it be clicked

Short answer

Google is predicting how often your ad will be clicked compared with the other ads competing for the same query. It is a prediction, not your measured click-through rate.

Two causes dominate, and the first is routinely missed. The first is copy: an ad that does not stand out, or that does not say why a searcher should click it. The second is that the ad is barely being shown at all — a low bid, a narrow schedule, tight targeting, or a low Ad Rank mean too few impressions for Google to form a confident estimate.

What to change

  • Check eligibility before you rewrite anything. An ad that is not being served cannot be clicked, and improving copy for it does nothing.
  • Write copy a competitor would not write. Specific and plain beats clever.
  • Test more than one ad per group. Two ads is a starting point, and the loser is information.
  • Avoid gimmicks: excessive capitals, punctuation stacks, and clickbait. They can lift a click-through rate and lose the conversion, which is the opposite of the point.

An estimate needs enough data. On a keyword with a handful of impressions, the rating is close to meaningless, and no amount of editing will change it quickly.

Why your high Quality Score campaign is not profitable

Because Quality Score measures relevance, and relevance is not revenue. The most common version of this story is a well-organised account with excellent ad relevance, fast pages and tidy ad groups, spending money on searches that do not want to buy. Every quality signal is satisfied. Nobody buys.

The reverse is just as common, and just as expensive: a single-word brand keyword with a low Quality Score that converts at twelve times the account average. The score says the ad is not compelling to a searcher who already intended to buy, which is true, and irrelevant.

The distinction that resolves both

Quality Score answers: does this impression look like a good match?

Your conversion data answers: did this searcher buy?

Only the second is evidence about money. Quality Score is evidence about a mismatch, which is useful — a low score is telling you the ad and the page are not landing where the searcher expected. It is not telling you whether the customer pays.

This is why every tool in this suite scores against targets the operator sets, and why none of them claims to predict a result. A metric that cannot be checked against outcomes should not be given a score of its own.

Why your low Quality Score keyword makes money

Because the keyword already converts, and the score describes something narrower than its apparent importance. Low ad relevance on a branded search term is close to normal: the searcher already knows who you are, the ad copy is not what made the decision, and the estimate is low because the ad is not doing the persuasive work.

Three cases where low is fine

  • Brand and navigational terms. The score is rarely high and the conversion rate is rarely anything else. Improving it is expensive and pointless.
  • Keywords with a proven return. If the economics work at the current cost per click, a better score that raises the price is a downgrade.
  • Keywords you are testing. A low score on a new term is a fact about the early data, not a verdict. Judge it on conversions when it has some.

And two where it is not

  • A keyword that should be your best and is not. High volume, high intent, a low score, and no conversions. That is a genuine mismatch worth fixing.
  • A page that loads badly. This is one of the few cases where the score is pointing at a real, measurable problem: people are leaving before the page arrives.

The rule this tool applies

Fix the score where the keyword matters commercially and the score disagrees with the result. Leave it alone everywhere else. Quality Score is a diagnostic to be read alongside your conversion data, never a target to be reached on its own.

Search and Display do not use the same words

Copying a Display diagnostic into a Search report is a common way to misread your own account. The two networks describe the same underlying assessments on different scales.

Search campaigns

Ad relevance, expected click-through rate and landing page experience are reported as above average or below average. There is no five-step scale and no “very low” in a Search keyword report.

Display and Video campaigns

The same three assessments are reported on a five-step scale: very low, low, below average, above average, high. “High” exists in Display and cannot be produced in Search, which is why a Display score of 8 feels better than a Search score of 8 and means nothing by comparison.

Placements, not just campaigns

In Display and Video, Quality Score is calculated per placement group, and only for placements where it applies. Two placements in the same ad group can have different scores and different reasons. If you are comparing a Display score with a Search score, you are comparing answers to two different questions.

This tool asks which network you are in, because the same word — “below average” — carries a different amount of information depending on the scale behind it.

Where Quality Score appears at all

Not everywhere, and a missing score is usually an eligibility problem rather than a bad one:

  • In Search, it sits in the keyword status column and needs enough impressions to estimate.
  • Smart Bidding needs a conversion history in the campaign before it will run, which is a separate requirement from having a score.
  • A paused keyword, or one that has never been eligible to serve, has none.
  • In Display and Video it is per placement group, and only where it applies.
  • Shopping and some campaign types do not show it in the same way at all.

Quality Score is not the page experience score

They overlap and are often confused. Page experience is about speed, responsiveness and visual stability. Quality Score's landing page experience factor is about relevance to the ad and the query, with load time as one part of it. Google recommends Core Web Vitals as the way to measure page experience, and Quality Score is not a Core Web Vitals score, despite how often the two are presented as the same number.

If your score is missing

Check that the ad is genuinely eligible to serve — schedule, bid, targeting, and budget — before optimising anything. Most missing scores are a delivery problem, and no rewrite fixes a delivery problem.

What to change, in the order worth doing it

The order is by value against effort, not by how alarming the rating looks. Page speed is last because it is genuinely slow to change, and because a slow page is blamed for a great many scores that were caused by something else.

  1. Ad group structure. One theme per ad group. This is free, it makes every other change easier, and it is the precondition for the rest of the list.
  2. Ad copy. Put the query's subject in the headline, answer the intent in the description, and add variants. This moves ad relevance and expected click-through rate together, which is why it comes first.
  3. Landing page relevance. A page specific to the ad, not a homepage. The promise in the ad has to be visible on the page, and the page has to be about this.
  4. Delivery checks. Schedule, budget, bid and targeting. A great ad that is not eligible to show will never look great.
  5. Page speed, last. Worth doing on its own merits, and often worth doing on mobile first. But fix relevance before speed, because a fast page that answers the wrong question still answers the wrong question.

How long it takes

Quality Score is calculated on historical signals, so it moves only after a change has gathered enough new impressions to measure. Expect weeks, not days, and expect the score to lag the change that caused it. Anyone promising a specific score inside a day is promising a number the platform has not computed.

How to tell whether a change worked

Change one thing at a time. Ad relevance and expected click-through rate are tightly linked, so rewriting both at once tells you which edit was the useful one only by luck.

  • Make the copy change, then wait for enough impressions before reading the score again.
  • Keep your own conversion rate and cost per conversion visible during the wait. A Quality Score improvement that arrives alongside worse economics is not a success.
  • Use the ROAS calculator to check whether the clicks you paid for are paying back, and the account audit scorecard to keep the account-level picture in one place.

What not to do

  • Do not restructure an account that already converts well. A profitable account with a low score is not broken.
  • Do not pause keywords purely on a low score. Check whether they convert first.
  • Do not chase a specific number across hundreds of keywords. The aggregate will not move, and the work will not be repeatable.

Traps worth naming

  • Reading the score as a forecast. It is not one, and Google says so directly.
  • Rewriting profitable brand ads. A low score there is normal and the work is wasted spend on effort.
  • Judging on a handful of impressions. The estimate is meaningless at low volume, and it will change on its own.
  • Blaming page speed first. It is the slowest lever and a frequent false culprit for a relevance problem.
  • Copying Display wording into Search. “Very low” does not exist in a Search report, so the scale has to be identified before it is compared.
  • Changing three things at once. Then nothing is learned and the same mistake gets made again.

What this tool will not tell you

  • It will not predict your cost per click, conversion rate or return from a Quality Score. Not because the information is hard to get, but because Quality Score does not contain it.
  • It will not tell you your Ad Rank. Ad Rank is the auction result, and its inputs and weights are not published.
  • It will not tell you whether a keyword is profitable. Only your own conversion and margin figures can, and the ROAS calculator is the tool for that.
  • It will not tell you what score to aim for. An account full of brand terms does not need one.
  • It will not connect to your account, read it, or change anything.
On Ad Rank. It is often described as bids multiplied by Quality Score, expected click-through rate and other factors. Treat that as a teaching model rather than a formula: the real calculation and the weight of each part are not published, and anyone claiming to compute your Ad Rank is guessing with your money.

What happens to your numbers

Short answer

Nothing leaves your browser. A single Quality Score and three dropdowns are not worth transmitting, so nothing is. Unlocking sends your name and email so the copy can be emailed to you — not your account data.

There is no account connection on this page, and there will not be one. Reading your real Quality Score would require access to your Ads account, and a tool that asks for that access to show you a number you can read in a table has not earned it.

Details worth stating plainly

  • No account data is requested, transmitted or stored, in either mode.
  • No cookies are set by this tool and no third-party script is loaded.
  • Unlocking stores a short-lived local token so you are not asked twice on the same device.
  • The emailed copy is the one place this diagnostic exists outside your machine. Ask and we will delete it.

Last reviewed 2026.

Quality Score questions, answered properly

Does a high Quality Score mean my campaign will perform well?

No. Google describes Quality Score as a diagnostic of predicted relevance, and states plainly that it does not predict performance or cost. A Quality Score of 10 does not mean a campaign makes money, and a Quality Score of 3 does not mean one will lose it. It measures how closely your ad and landing page match the query, which is a different question from whether the customer buys.

Is a low Quality Score always a problem that needs fixing?

No, and this is the most commercially important thing to know about Quality Score. If a keyword already converts profitably at a low Quality Score, there may be nothing to fix, because the returns are real. Raising Quality Score usually costs more per click, because a better-matched ad tends to be clicked more in a competitive auction. On brand terms, low Quality Score is common and usually harmless. Chasing a higher score across the whole account can cost more than the improvement is worth.

What are the three factors in Quality Score?

Ad relevance, expected click-through rate, and landing page experience. Ad relevance asks whether your ad text is relevant to the keyword. Expected click-through rate is Google's estimate of how often your ad will be clicked compared with competing ads for the same keyword. Landing page experience asks whether the page you land on is relevant to the ad, and whether it loads and behaves well, particularly on mobile. Google shows each one separately, and each has its own specific causes and its own fix.

Why is my Quality Score not shown?

There are several ordinary reasons. A keyword needs enough data before Google will estimate it, and an ad that is barely being shown has too few impressions. Smart Bidding also requires a conversion history in the campaign before it will run. In Display and Video campaigns Quality Score is only calculated for placements where it applies. A keyword that is paused, or that has never been eligible to serve, will not have one. If your score is genuinely missing, check that the ad is actually eligible to serve before optimising anything else.

How do I raise Quality Score fast?

You cannot, and the expectation is the problem. Quality Score is calculated on historical signals, so it moves as a consequence of changes that have already gathered enough impressions to measure. Any tool promising a score of 10 in an afternoon is promising a number the platform has not computed yet. The reliable route is slower and unglamorous: tighten ad groups so each one is a single theme, make the headline say what the query wanted, write copy a real person would click, and fix the landing page so it delivers the promise the ad made.

Sources and method

What comes from Google

The four band edges — above 7 excellent, 4 to 7 good, 1 to 4 needs improvement, below 1 poor — and the three factors behind the score. The two diagnostic scales: above or below average for Search, and a five-step scale from very low to high for Display and Video. Also the central point of this tool: Google's own position that Quality Score is a diagnostic and does not predict performance.

Ad relevance and expected click-through rate are per keyword. Landing page experience covers relevance to the ad as well as how the page loads and behaves, and Google recommends Core Web Vitals for measuring page experience separately.

What is judgement, and whose

Documented platform behaviour: the band edges, the three factors, the two diagnostic scales, per-keyword and per-placement reporting, and the requirement for a conversion history before Smart Bidding will run.

Judgement on this page, and clearly marked as such: the order in which the fixes are worth doing, the observation that improving a Quality Score usually raises the price, and the recommendation to leave a profitable keyword alone. These are opinions about where effort pays off, not facts about the platform, and the tool labels the commercial consequence as a warning rather than presenting it as a rule.

Not asserted anywhere: the combination's weights, any Ad Rank figure, any predicted cost per click, conversion rate or return. All of these would require data this page does not have.

Last reviewed 2026. Platform behaviour changes; if your account says otherwise, trust your account over this page and tell us.