Key takeaways
- Google began testing “Strongest Match” and “Strong Match” ad relevance labels in the US on June 26, 2026. GrowByData’s scans caught them three days later, weeks before mainstream coverage caught up.
- Across 300,641 mobile scans in August 2026, a label showed up on just 1.65% of results, 4,966 scans total.
- When a label did appear, Strong Match showed up roughly twice as often as Strongest Match (67.5% vs. 32.5%), consistent with only one ad per auction qualifying for the top tier.
- Automotive accounts for 43% of every label observed, but that’s a volume effect, not a higher labeling rate. Its own text-ad label rate is a middling 2.5%.
On June 26, 2026, Google’s Ads product liaison Ginny Marvin did something Google rarely does: she flagged a test before anyone had caught it in the wild. She said Google would run a small experiment adding a “Strongest Match” or “Strong Match” label to some search ads in the US, meant to show which results Google’s own systems think best answer the query.
The labels don’t touch the auction. Ad Rank still decides which ads show and where. Only after that’s settled does Google compare the ads that made the cut and hand the most relevant one a “Strongest Match” badge, while others that clear a lower bar get “Strong Match.” Marvin has said it would be unusual for more than one ad in the same auction to earn the top label, which turns out to matter for why one badge shows up roughly twice as often as the other.
GrowByData’s mobile SERP scans picked up the labels on June 26th, three days after Marvin’s announcement. Search Engine Roundtable didn’t report seeing it live until August 7th, six weeks later. We spent that stretch watching a full month of August data to see what the test actually looks like at scale: 300,641 mobile scans, one industry benchmark at a time.
What are the “Strongest Match” and “Strong Match” labels?
They’re a two-tier relevance badge, applied after the auction closes. “Strongest Match” is Google’s single best guess at which ad most closely answers the query, drawn from whichever ads already survived the auction. “Strong Match” is the runner-up tier, still elevated, not the top pick. Google hasn’t published what specifically separates the two, or what keeps an ad from getting either label at all.
That gap matters. Nobody’s published the criteria for a badge that could shape how a user reads your ad, and there’s no dashboard telling advertisers whether they’ve earned it.
Strong Match is winning, by roughly two to one
Across the full month of August, GrowByData’s mobile scans covered 300,641 search results carrying text ads. A label showed up on 4,966 of them, 1.65% overall. When one did appear, it was Strong Match far more often than Strongest Match, and that’s not really a surprise once you know how the labels get assigned: only one ad per auction can walk away with the top badge, while a wider pool clears the bar for the runner-up.
| Metric | August 2026 (full month) |
|---|---|
| Total scans | 300,641 |
| Scans with any label | 4,966 (1.65%) |
| Strongest Match share of labels | 32.5% (1,612) |
| Strong Match share of labels | 67.5% (3,354) |
For two days in late August, label detection went completely dark across every account we track, real scan volume, zero labels of either kind. When it came back, only Strong Match resumed. Strongest Match hasn’t shown up since. That’s not the pattern a small-scale live experiment usually produces on its own, and we’re checking with our own data team whether it’s a real shift on Google’s end or a break in our tracking. We’ll correct this figure if it’s the latter.
Google hasn’t said anything public about recalibrating the criteria. Whatever’s behind the gap toward Strong Match, it’s not something you’d catch from one screenshot. It has to be tracked continuously.
Why might your own monitoring miss this entirely?
GrowByData only picked these labels up in incognito-session scans. Run the same keyword, same device, same day through a logged-in or manual session, and the label doesn’t show. That’s not a fluke. It’s a session-state test, not just a query or device test.
If your SERP monitoring runs exclusively through manual or logged-in scans, a test built this way passes right through it. You wouldn’t be behind on the story. You’d never see it at all.
If you’re not sure whether your current SERP monitoring would catch something built this way, get a visibility audit and we’ll show you what your setup does and doesn’t pick up.
Text ad labels by industry, August 2026
The labels weren’t spread evenly across industries, and there are two honest ways to look at that, depending on what you’re actually asking.
Which industries see these labels most often, relative to their own ad volume? By that measure, Food, Travel, and Furniture lead, each landing between roughly 2.5% and 2.7% of their own text ad scans. None of those are huge samples on their own, Food in particular is only 552 scans for the month, so treat the ranking as directional rather than definitive.
| Industry | Strong Match % | Strongest Match % | Combined |
|---|---|---|---|
| Food | 1.63% | 1.09% | 2.72% |
| Travel | 1.78% | 0.90% | 2.68% |
| Furniture | 1.49% | 1.04% | 2.53% |
| Automotive | 1.63% | 0.84% | 2.47% |
| Real Estate | 1.56% | 0.78% | 2.34% |
| Apparel | 1.45% | 0.61% | 2.06% |
| Brand | 1.60% | 0.39% | 1.99% |
| Technology | 1.16% | 0.61% | 1.77% |
| Software | 0.97% | 0.36% | 1.33% |
| Retail | 0.83% | 0.36% | 1.19% |
| Health | 0.70% | 0.30% | 1.00% |
| Marketing | 0.63% | 0.25% | 0.88% |
| Finance | 0.21% | 0.07% | 0.28% |
Flip the question, ask which industries account for the sheer volume of labels instead, and the picture changes. Automotive alone makes up 43% of every labeled listing across every industry we track, not because Automotive ads get labeled unusually often, its own rate above is a middling 2.5%, but because Automotive simply runs far more scanned ad volume than any other category we cover. Three industries we track, Entertainment, Insurance, and Manufacturing, saw no labels at all this month, though they also run comparatively little volume, so that’s not much of a finding either.
| Industry | Strong Match | Strongest Match | Combined share |
|---|---|---|---|
| Automotive | 28.2% | 14.6% | 42.7% |
| Retail | 10.8% | 4.6% | 15.4% |
| Health | 8.3% | 3.5% | 11.8% |
| Real Estate | 7.1% | 3.5% | 10.6% |
| Apparel | 3.9% | 1.7% | 5.6% |
| Technology | 2.9% | 1.6% | 4.5% |
| Travel | 2.9% | 1.5% | 4.4% |
| All remaining industries | 3.5% | 1.5% | 5.0% |
Should advertisers change anything today?
Not yet. Chasing an unconfirmed signal is usually how a quarter gets wasted.
Check whether your monitoring tells incognito sessions apart from logged-in ones. This is the specific test that exposed the gap, and it won’t be the last one built this way.
Keep working the fundamentals Google itself keeps pointing back to: ad relevance, how closely your copy matches the query, landing page experience. None of that changes whether the badge ever reaches your account, but it’s the same list Google was already grading you on before anyone could see the grade.
Watch mobile specifically. Every instance tracked so far is mobile-only, and a desktop-heavy QA process is the easiest way to miss it.
Still an open test
Google hasn’t committed to expanding this, and it hasn’t ruled it out either. What’s certain: Strong Match is the more common label right now, there’s a real structural reason to expect that, and there’s also a real question mark hanging over the back half of our August data that we haven’t closed out. We’ll keep watching both.
A quick methodology note: this is 300,641 US mobile SERP scans from GrowByData’s SERP Watch dataset. Industry totals exclude a small unclassified bucket, and Google has described the underlying test as limited to a small share of US users. Treat these numbers as a real, large sample of what we saw, not a census of Search traffic overall.
Curious whether your own ad accounts are currently showing these labels, or your competitors’ are? Talk to one of our search intelligence analysts and we’ll pull your live data.
