The platform states that the word-for-word phrase from a search term does not need to be on your landing page. It also states that its quality score is not a key performance indicator, should not be optimized, and is not an input in the ad auction.
Both of those sentences are on its own help pages. Both contradict how the industry teaches this subject. And the practice they are used to justify, keeping your page consistent with your ad, is still good practice, for reasons that have nothing to do with either.
This page separates the documented mechanism from the folklore around it, and says what the evidence actually supports.
What the score is, and what the platform says about using it
Start here, because almost every conversation about message match is really a conversation about this number.
The definition. A diagnostic tool meant to give you a sense of how well your ad quality compares to other advertisers. Measured 1 to 10, available at keyword level.
The three components. Expected clickthrough rate, the likelihood that your ad will be clicked when shown. Ad relevance, how closely your ad matches the intent behind a user’s search. Landing page experience, how relevant and useful your landing page is to people who click your ad.
How each is reported. Above average, average or below average, compared with other advertisers on the same exact search over the past 90 days.
The sentence that should end most agency arguments. The platform states that quality score is not a key performance indicator and should not be optimized or aggregated with the rest of your data.
And the one that should end the rest of them. It states that quality score is not an input in the ad auction, and that it is a diagnostic tool to identify how ads showing for certain keywords affect the user experience. Elsewhere: use it to diagnose low quality ads, rather than trying to optimize the 1 to 10 score you receive.
What follows. Reporting an account’s average quality score to a client is reporting a number the vendor has explicitly told you not to aggregate.
The Ad Rank formula is not a formula
The multiplication everyone writes on whiteboards does not appear in the documentation.
What the platform publishes. Ad Rank is described as a set of values used to determine whether ads are eligible to show and, if eligible, where on the page.
The six factors, in full. Your bid. The quality of your ads and landing page. The Ad Rank thresholds. The competitiveness of an auction. The context of the person’s search. The expected impact from your ad assets and other ad formats.
What is absent. Any arithmetic. No multiplication, no weighting, no function. “Ad Rank equals bid times quality score” is industry shorthand that the platform has never written.
What the quality factor actually says. That the platform assesses the usefulness and relevance of your ad and landing page, the expectations users have about your landing page based on the clicked ad creative, ease of landing page navigation, and other search quality signals.
Read that phrase again. The expectations users have about your landing page based on the clicked ad creative. That is the real mechanism, and it is about expectation, not vocabulary.
Why the distinction matters practically. An ad promising a free trial that lands on a page selling an annual contract fails this regardless of how many words the two share. An ad promising a free trial landing on a page that delivers one passes it regardless of how few.
It is not silent on the relationship. It is just asking for something other than word matching.
On the landing page itself. Choose a landing page that closely matches your ad and keywords. The worked example given: if your keyword is discount shoes and your ad promises shoes at 20% off, customers should be able to find and buy shoes at that discounted price on your landing page.
On the call to action. Your landing page should mirror the call to action in your ad text. The platform describes the connection between ad and landing page as the bridge between a potential customer and a purchase.
Elsewhere, more plainly. Keep messaging consistent from ad to landing page, and ensure the page follows through on the ad’s offer or call to action.
Note the verbs throughout. Deliver. Follow through. Find and buy. Mirror the call to action. Every one of them is about the promise being kept, not about the sentence being repeated.
What the historical guidance listed, on a page that has since been merged away: offer relevant, useful and original content; promote transparency and foster trustworthiness by explaining what your business does before asking visitors to fill out forms and making contact information easy to find; make navigation easy on mobile and desktop without pop-ups that interfere; and decrease loading time.
The three-pillar summary that agencies still quote. Relevance, transparency and navigability. It is accurate as a summary of the older page, and worth knowing that the current documentation no longer frames it that way.
Its documentation says the thing people wish the larger platform said, and adds a mechanical definition worth knowing.
Same shape. A quality score from 1 to 10, with expected click-through rate, ad relevance and landing page experience as components.
A blunter disclaimer. Its documentation states that quality score is an evaluation based on historical performance in past auctions, that it is not used at auction time to determine ad rank, and that it does not affect cost or spend.
A window it actually discloses. Its score is based on search queries and other input that matched keywords exactly in the last 30 days. The larger platform’s equivalent is 90 days for component comparison, but it does not describe its calculation window this way.
And a landing page definition that is almost operational. It describes the landing page experience score as based on how many times customers leave your ad’s landing page shortly after arriving.
Why that last one is useful to know. It tells you what at least one platform believes a bad landing page experience looks like from the outside: fast departures. That is a measurable thing you can look at yourself.
The caution attached. That is one platform’s stated approach, not a general law, and it should not be assumed to describe how the larger platform computes anything.
The social platform does not do this at all
Worth stating, because people carry search habits across and look for a score that is not there.
What exists. Relevance diagnostics, which replaced an older 1 to 10 relevance score. They assess an ad’s past performance in the auction over your selected date range.
The three rankings. Quality ranking, how the ad’s perceived quality compared with ads competing for the same audience. Engagement rate ranking, how its expected engagement rate compared. Conversion rate ranking, how its expected conversion rate compared with ads sharing the same optimisation goal.
What quality ranking measures. The ad, not the page. Signals include feedback from people viewing or hiding the ad, and assessments of low-quality attributes in the ad such as withholding information, sensationalised language and engagement bait.
The threshold. Diagnostics are not available for ads with fewer than 500 impressions.
And the same disclaimer. The relevance diagnostics are not inputs into the ad auction.
What is missing. Any structured landing page experience score. The platform mentions post-click experiences in its advertising policies, but there is no scored, separately reported landing page component of the kind the search platforms publish.
Where the term came from, and what has actually been measured
The concept is popular, well-taught and, on its literal reading, unevidenced.
The definition in circulation. A measure of how well your landing page copy matches the phrasing of the ad or link that brought the visitor there, which for search advertisers means matching ad copy to the landing page headline. The associated line, from the software company that popularised it: if your ad copy does not match the message on your landing page, you are disrespecting the click.
Who coined it, and when. I could not establish either. The glossary defining it claims no date and no authorship, so “popularised by” is the accurate phrasing, not “coined in”.
The figure everyone quotes. An agency experiment run in mid-2016 reported a conversion rate moving from 0.49% to 0.73%, a 47.6% relative increase, on a required sample of 9,488 against an actual sample of 239,607, at above 95% confidence. That is a genuinely well-powered test and it deserves credit for publishing its sample.
What that test actually changed. The imagery. It tested repeating the same photograph between the ad and the page, not matching the wording. It is evidence for visual continuity, and it is regularly cited as evidence for copy matching, which it is not.
The two real academic studies, and what they measure. A 2009 paper in a top management journal analysed a six-month panel of 9,664 observations across 1,878 unique keywords from one large national retailer, and reported that a one unit increase in the advertiser’s landing page quality score is associated with a conversion rate increase of as much as 22.5%, with cost per click negatively associated with landing page quality.
The caveat that paper states about itself. The platform did not disclose the score, so the authors reconstructed it: two independent annotators rated each landing page on relevance, transparency and navigability, with an interrater reliability of 0.73. It is a study of overall page quality, not of textual correspondence.
The second one. A 2009 conference paper from inside the search company itself found that ads following its published landing page guidelines had a mean bounce rate 25.4% lower than those that did not, significant at p below 0.0001, on a sample of a few thousand expert-evaluated ads, with a separate experiment on 7,000 ads weighted by impressions across eight languages.
The honest summary. Nobody has isolated literal text matching in a published study. What has been measured is the broader construct of landing page quality, and it does show up in conversion rate and in cost per click.
Six things, and none of them requires optimising a score the vendor has told you not to optimise.
Keep the promise, not the words. If the ad offers a free trial, the page delivers a free trial, visibly, above the fold. Whether the headline repeats the phrasing is a stylistic choice.
Make the offer findable in one screen. The platform’s own worked example is a visitor being able to find and buy the discounted product. If your promised thing requires scrolling and a menu, you have failed the standard that is actually documented.
Mirror the call to action. This one is explicit in the guidance: the action promised in the ad should be the action available on the page. An ad saying “book a demo” landing on a page whose only action is a newsletter signup is the clearest possible failure.
Watch fast departures, not score. One platform tells you plainly that it looks at how many people leave shortly after arriving. That is something you can measure on your own account today.
Fix the boring things. Load time, mobile navigation, no interstitials, and an explanation of what your business does before you ask for a form. That list is from the platform’s own historical guidance and it has not stopped being true.
Stop reporting average quality score. It is a diagnostic. Use it to find the keywords with a below-average landing page component, then go and look at those pages.
The platform says the words do not have to match. Its guidance states the word-for-word phrase from a search term does not need to be on your landing page.
It also says the score is not for optimising. Quality Score “is not a key performance indicator and should not be optimized or aggregated with the rest of your data”.
And that it is not an auction input. Both search platforms say this; one adds that it “does not affect cost or spend”.
The Ad Rank formula is folklore. The documentation lists six qualitative factors and publishes no arithmetic.
What is documented is expectation: the expectations users have about your landing page based on the clicked ad creative.
One platform gives an operational definition: how many times customers leave your ad’s landing page shortly after arriving.
The largest social platform does not score landing pages at all, and gives no diagnostics below 500 impressions.
The 47.6% figure everyone cites tested imagery, not copy. The two real academic studies measure overall landing page quality, finding up to 22.5% conversion association and 25.4% lower bounce rate.
Does my landing page have to repeat my ad headline word for word?
No, and the platform says so directly. Its guidance states that the word-for-word phrase from a search term does not need to be on your landing page, and directs you instead to deliver what the user is looking for.
Is Quality Score worth optimising?
The platform says not. It states that Quality Score is not a key performance indicator, should not be optimized or aggregated with the rest of your data, and is not an input in the ad auction. It is a diagnostic.
Is Ad Rank bid multiplied by Quality Score?
That formula is not a platform statement. The documentation lists six qualitative factors: your bid, the quality of your ads and landing page, the Ad Rank thresholds, auction competitiveness, the context of the search, and the expected impact of assets and formats.
What does landing page experience actually cover?
Relevant, useful and original content; transparency and trustworthiness about your business; easy navigation on mobile and desktop; and fast loading. The historical framing was relevance, transparency and navigability.
How does the second search platform define it?
More mechanically. It defines the landing page experience score as based on how many times customers leave your ad's landing page shortly after arriving, and computes its quality score on exact keyword matches from the last 30 days.
Does the largest social platform score landing pages?
Not in the same way. Its relevance diagnostics rank quality, engagement rate and conversion rate, and its quality ranking is about the ad itself, measured through signals like people hiding it and low-quality attributes in the ad.
Is there any study proving message match works?
Not for literal text matching. The two academic studies in this area measure overall landing page quality, and the widely cited 47.6% figure comes from an agency test of visual imagery consistency, not of wording.
So should I ignore the advice?
No. Keeping the promise you made is sound for reasons that need no study: a visitor who cannot find what they clicked for leaves. Just stop justifying it with a score the platform tells you not to optimise.