Why was my Google ad disapproved? Unreliable claims, misrepresentation, and destination mismatches

Short answer: a Google ad is disapproved because a named policy failed on the RSA, an asset, or the destination — not because “Google is picky.” The three labels that eat the most search and Performance Max spend are unreliable claims, misrepresentation, and destination mismatch. They are different fixes. Mixing them up is why people rewrite headlines for a tracking-URL problem, or add “results vary” under a miracle claim and still sit in Limited.

This is not another 20-reason dump. We map those three labels to Google’s published pages, then score three Responsive Search Ads the way review actually works: every headline, every description, sitelinks, and the final URL. Meta’s personal-attributes rejections are a different system; if you paste the same primary text into Google, you will not get the same fail. We covered that split in same copy, three policy systems.

Read the label before you rewrite

In Google Ads: Campaigns → Ads & assets → Ads. Filter by Policy: Disapproved (or Limited). Open the ad → View policy details. The string you see is a pointer into Google Ads policies, not a suggestion.

Unreliable claims is not a sibling of misrepresentation. It is a named subsection of Misrepresentation. The UI may show either label for the same sentence. Destination mismatch is a different family: Destination requirements / destination mismatch. Copy edits will not fix a cross-domain redirect.

Google also has nearby labels — unclear relevance, unavailable offers, destination experience, insufficient original content. Those sit under the same two policy trees. Treat them as siblings, not a fourth mystery.

Google reviews the RSA as a pile of claims, not one paragraph

A Responsive Search Ad can pin up to 15 headlines and 4 descriptions. Review sees combinations. If headline 11 still says “Guaranteed 15 lbs,” it does not matter that headlines 1–3 are tame. Sitelinks, callouts, and structured snippets are assets too. A sitelink “70% off everything” is an offer Google will look for on the destination.

YouTube, Demand Gen, and Display add extra creative rules (misleading before/after imagery, sensational health headlines). The three labels below still apply. Search RSAs are where most advertisers first meet them.

Three ads, three labels

Each sample is the kind of RSA a performance team actually ships. “Likely disapproved” means the published policy text describes that pattern. It is not a pass-rate. Restricted verticals (healthcare, financial products, crypto) can still fail after a rewrite if the account lacks certification.

Ad 1 — unreliable claims: Google’s own weight-loss example, in RSA form

Google’s unreliable claims page is unusually concrete. Under health and weight loss it disallows “claims of unrealistic weight loss within a specific time frame or requiring little effort.” The documented example is an ad that says you can eat whatever you want and lose 10 pounds in a month. Headline 1 is the timeframe. Headline 2 is the little-effort clause. Together they are that example, not a close cousin.

“Guaranteed results” does not save it. The same page says: if you guarantee certain results, you must have a clear, easily accessible refund policy; testimonials that claim specific results must show that results can vary. That is an extra duty when you use a guarantee. It is not permission to keep an improbable likely-outcome. “Clients drop a size every week” is a typical-result testimonial without the required disclaimer.

Same policy, money version (so finance advertisers do not bounce): Google also disallows “unrealistic promises of large financial return with minimal risk, effort or investment,” including guaranteeing returns and presenting products as risk-free. Swap the headlines for “12% monthly, zero risk” and you are on the financial examples on the same page.

What to change (not just the disclaimer)

The rewrite describes the product and points at terms. It drops the lbs/days number, the “no gym” miracle, and the implied typical weekly loss. Put the refund policy on that URL, above the fold or one obvious click away — Google’s words are “clear and easily accessible,” not “buried in a 4,000-word footer.”

If the landing page still leads with “15 lbs / 14 days” in the H1 or a testimonial carousel with no variance disclaimer, the RSA rewrite is wasted. Review includes the destination.

Ad 2 — misrepresentation: the 70% off that is not on the page

What is actually on /: a homepage hero for new-in denim, a 20% banner on one collection, no end date, and /collections/shoes at full price. The RSA can be grammatically true in the marketer’s head (“we meant the promo landing page we did not use”). Review does not use that head.

Misrepresentation lists several clauses that fire together here:

• Unavailable offers — promising products or promotions in the ad that are unavailable or not easily found from the destination.
• Unclear relevance — the promotion is not relevant to the destination (homepage denim vs “70% off shoes”).
• Dishonest pricing (same policy family) — a sitewide 70% that is really 20% on a subset.

This is why “we made the headlines more compliant” fails. The headlines are the problem relative to that URL. Two legal fixes, pick one:

1. Point the final URL at a sale page that shows 70% off the advertised set, with the Sunday midnight end time visible, and make the shoes sitelink land on shoes that are actually 70% off.
2. Rewrite the RSA to the offer that exists: “20% off the Fall denim collection — see dates on site.”

Check callouts and promotion assets the same way. A callout “No exclusions” against a page full of exclusions is the same unavailable-offer finding. Open the final URL in an incognito window, on a phone, not logged into Shopify preview.

Ad 3 — destination mismatch: the copy is fine, the hop is not

Google’s destination mismatch page is operational, not rhetorical. AdsBot visits the final URL and compares it to what the ad claims. Documented fail modes that match this RSA:

• Display URL domain northline.example vs a crawled destination on aff-track.net or northline.co.uk.
• Redirects from the final URL that take the user to a different domain.
• URL shorteners whose expanded URL is not the final URL you declared.
• Geo-redirection that sends crawlers or users to another country domain than the one in the ad.
• Case-only path mismatches (/Shred vs /shred) on case-sensitive servers.

If you must pass a dynamic click ID on the final URL, Google’s tracking docs say to put {ignore} immediately before those parameters so AdsBot can ignore the junk and still see a stable page — see ValueTrack / {ignore} and the mismatch help article’s own example. Uncertified third-party click trackers are called out as a disapproval risk. Use a Google-certified tracker or first-party redirects that stay on the same domain.

Related labels that are not mismatch but get filed with it: destination experience (interstitial popups, broken mobile layout) and destinations that block AdsBot-Google. If AdsBot cannot crawl, Google cannot confirm the page matches the ad. Allowlist the bot; do not put the offer behind a hard login wall.

What these three labels are not

Do not burn a week on headline tone if Policy details says something else. Common mix-ups:

• Healthcare and medicines / unapproved substances — the SKU itself is restricted. A quieter headline will not certify an unapproved supplement.
• Financial products and services — many products need country-specific disclosures and account certification. Unreliable claims can stack on top.
• Trademarks — using another brand in the RSA is a different policy.
• Editorial and professional requirements — spelling, gimmicky capitalization, and “FREE!!!!” punctuation. Easy to confuse with misrepresentation; the fix is copy hygiene, not a new landing page.

How to fix a Google Ads disapproval

1. Open Ads & assets, select the disapproved ad, and read View policy details. Note the exact label (unreliable claims, misrepresentation, destination mismatch, or another policy).
2. If the label is unreliable claims or misrepresentation, rewrite RSA headlines and descriptions so they describe the product, price, or process — not a guaranteed, overnight, or risk-free outcome. Match every sitelink to a real page.
3. Open the final URL logged out, on mobile. Confirm the offer, price, dates, and product in the ad are visible without a login or a different geo-domain.
4. If the label is destination mismatch, set final URL to the real HTTPS page, make the display-URL domain match, remove cross-domain shorteners, and prefix dynamic click IDs with {ignore}.
5. Save the edited asset and wait for automatic re-review (typically 24–48 hours). Appeal only after the change, and only if you think the new version is compliant.

Cloning the ad so “a fresh one can pass” is how accounts pick up repeated violations. Edit the asset Google already flagged. If several RSAs share a sitelink or promotion asset, fix the shared asset once or every ad that uses it stays dirty.

A 10-minute preflight before you hit Enable

1. Scan every RSA headline and description for timeframe body results, “guaranteed,” “risk-free,” “#1,” overnight, and “no effort.” Delete or replace — do not hide them in headline 14.
2. If any line is a testimonial with a specific result, the destination must show that results vary (visible, not footer-only).
3. If any line is a guarantee, the refund policy must be obvious on that same URL.
4. Click every sitelink and the final URL logged out on mobile. Offer, price, and dates must be findable without guessing the campaign name.
5. Confirm display-URL domain === final-URL domain after redirects. No bit.ly, no affiliate domain in the crawl, no geo hop to a different ccTLD.
6. Prefix dynamic click IDs with {ignore}. Confirm AdsBot-Google is not blocked.

Frequently asked questions

Sources

• Google Ads policies (index)
• Misrepresentation
• Unreliable claims
• Destination requirements
• Destination mismatch
• ValueTrack parameters and {ignore}

Policies change. This is a reading of the public Google Ads policy pages as of September 14, 2026, not a guarantee of approval and not legal advice. Restricted categories can still fail a clean RSA. Use the live account’s Policy Manager as source of truth, then resubmit.