Your Rankings Moved on 1 August and Google Confirmed Nothing. Here Is How to Diagnose It.
Trackers spiked from 1 to 6 August 2026 while Google logged no update. Four causes draw that chart. Here is how to tell them apart before you change anything.
Third-party rank trackers moved sharply from 1 August 2026, peaked around 3 August and flared again on 5 and 6 August. Google has confirmed no update at all.
- The dashboard is empty: The Google Search Status Dashboard records no ranking update and no indexing, crawling or serving incident for any day between 1 and 6 August 2026. The last confirmed ranking change is still the June 2026 spam update, logged at 2 days 1 hour from 24 June.
- The reports do not agree with each other: Publishers described Search rankings, Google Discover, indexing and analytics all misbehaving in the same week. That combination is unlikely to have one cause, which is the strongest argument against reading it as a single algorithm update.
- Four causes draw the same chart: A ranking change, a serving bug, a reporting artefact and your own last deploy all look identical in a traffic graph. They stop looking identical the moment you check average position, compare surfaces, and confirm the drop in a source Google does not compute.
- There is a recent precedent for guessing wrong: On 3 April 2026 Google said a logging error had been inflating Search Console impressions since 13 May 2025, roughly 50 weeks. Clicks were never affected, so every CTR calculated over that window was lower than reality, and the drop when the fix landed was a correction rather than a loss.
- Do the cheap diagnosis first: Three of the four causes are ruled out with data you already hold. Only rush a fix when position genuinely moved, the loss clusters on pages that share a pattern, and a second data source agrees that fewer people arrived.
An unconfirmed update is not a licence to rewrite the site. It is a reason to find out which of four very different problems you actually have.
Deciding whether a drop is a ranking change, a bug or your own last deploy needs Search Console, analytics and a crawl in the same view. CrawlRaven joins all three into one ranked plan, from $49 at launch. Try CrawlRaven free: 1 site, no credit card →
On 1 August 2026 the rank trackers lit up. By 3 August the spike was showing on nearly every public volatility index, and the forums filled with the usual question in the usual tone: what did Google just do?
Google's answer, as of today, is that Google did not announce anything. The Search Status Dashboard records no ranking update and no incident for any day in that window.
That gap between what people are measuring and what Google has confirmed is where the damage usually happens. So this piece is not a recap. It is the diagnostic I run before touching anything:
- What is actually established about early August, and what is not.
- The four very different causes that produce an identical-looking traffic chart.
- The order to test them in, cheapest and most reversible first.
What actually happened between 1 and 6 August
Strip out the speculation and a short list of facts survives. Volatility began on 1 August and peaked around 3 August, with spikes reported across AccuRanker, Algoroo, AWR, CognitiveSEO, DataForSEO, Mangools, Mozcast, Semrush and Serpstat.
It then flared again on 5 and 6 August. Alongside the ranking movement, publishers reported problems that have nothing to do with ranking at all.
- Abrupt traffic losses with no warning and no obvious pattern by topic or page type.
- Unstable Discover behaviour, which is a different surface with its own eligibility rules.
- Indexing oddities, which sit upstream of ranking rather than inside it.
- Analytics disruption, which is a reporting layer and not a search behaviour at all.
What moved in early August 2026
Six days of volatility against an empty status dashboard
The last ranking change Google has confirmed. Started 24 June, finished 26 June, logged at 2 days 1 hour. Global, all languages.
Rank-tracking tools register a sharp burst of reshuffling. No announcement from Google accompanies it.
Spikes reported across AccuRanker, Algoroo, AWR, CognitiveSEO, DataForSEO, Mangools, Mozcast, Semrush and Serpstat.
Volatility heats up again. Publisher reports now mix Search rankings with Discover, indexing and analytics problems.
No ranking update, and no indexing, crawling or serving incident is recorded for any day in the window.
Update dates and durations from the Google Search Status Dashboard. Tracker names as reported by the SEO trade press.
Read that list again, because it is the most useful thing in this article. Ranking, Discover eligibility, indexing and analytics are four different systems.
A single algorithm update does not break all four at once. The most likely reading is several unrelated things happening in the same week and being narrated as one event.
An "unconfirmed update" is not a category Google recognises. It is what the industry calls a period when trackers move and Google says nothing. Sometimes it was a real ranking change. Sometimes it was a bug, a test, or several small things at once.
What Google has confirmed, which is nothing
The Search Status Dashboard is the only authoritative record of what Google has announced. Here is everything on it for 2026, up to 9 August.
Every ranking change Google confirmed in 2026
| Entry | Started | Duration | Type |
|---|---|---|---|
| February 2026 Discover update | 5 Feb 2026 | 21 days 17 hrs | Discover |
| Serving issue | 24 Feb 2026 | 15 mins | Incident |
| March 2026 spam update | 24 Mar 2026 | 19 hrs 30 mins | Spam |
| March 2026 core update | 27 Mar 2026 | 12 days 4 hrs | Core |
| May 2026 core update | 21 May 2026 | 11 days 21 hrs | Core |
| June 2026 spam update | 24 Jun 2026 | 2 days 1 hr | Spam |
| July 2026 | Nothing recorded | – | – |
| August 2026, to 6 Aug | Nothing recorded | – | – |
Read off the Google Search Status Dashboard on 9 August 2026.
The last confirmed ranking change is the June 2026 spam update, which started on 24 June and finished on 26 June, logged at 2 days 1 hour and applied globally across all languages. Nothing has been recorded since.
Before that, 2026 has been busy: a February Discover update at 21 days 17 hours, a 15-minute serving issue on 24 February, a March spam update at 19 hours 30 minutes, a March core update at 12 days 4 hours, and a May core update at 11 days 21 hours.
Now the important caveat, because a dashboard is easy to over-read.
- It lists announcements, not changes. Google ships changes to Search constantly and announces a small fraction of them.
- An empty dashboard is not an alibi. It means nothing was announced, not that nothing happened to you.
- It still goes first. A named, dated update is the one explanation you can confirm rather than infer, so rule it in or out before you reach for a theory.
Four causes that draw the same chart
Here is the trap. A line that falls off a shelf on 1 August looks identical whether the cause was Google, a bug, a reporting error or your own release.
The chart cannot separate them. Four other things can, and none of them requires waiting for an announcement.
Four causes that draw the same chart
Impressions and clicks fall together, average position worsens, and the losses cluster on pages that share a template, a topic or a quality problem.
Average position is flat. If you still rank where you ranked, the ranking did not change.
Sample the lost queries and check who replaced you in the results.
One surface falls off a cliff in a single day while the others carry on. Discover is the usual victim, and position on Search never moves.
The decline is gradual, or every surface moves by a similar proportion on the same day.
Compare Search, Discover and News side by side before changing anything.
The chart moved and the business did not. Sessions, conversions and revenue hold steady while an impression or position metric steps to a new level.
Server logs, GA4 and your order data agree that fewer people arrived.
Confirm the drop in a second data source that Google does not compute.
The break lines up with a deploy, a migration or a plugin update, and it hits one section, one template or one URL pattern rather than the whole site.
Nothing changed on the site in the fortnight before the drop, on any system.
Crawl the affected URLs and diff robots rules, canonicals, noindex and redirects.
Work top to bottom. Three of the four are ruled out with data you already have.
Three of those four are ruled out with data you already hold, in under an hour, and none of the checks changes anything on the site. That is the entire argument for doing the diagnosis before the fix.
The diagnostic, in the order that costs you least
Order matters here. Run the checks that cannot break anything, and that can end the investigation outright, before the ones that cost a week of engineering time.
Did impressions move, or only clicks?
This single split does more diagnostic work than anything else in Search Console, because impressions and clicks fail for different reasons.
- Impressions down, position worse. You are being shown less often because you rank lower. This is the ranking-change shape.
- Impressions steady, clicks down, position steady. You still rank, and fewer people are clicking. Look at what changed on the results page above you, not at your rankings.
- Impressions down, position steady. Demand fell, or you lost eligibility for a surface or a feature. Check seasonality before anything else.
- Everything down, uniformly, overnight. Uniformity is the signature of a reporting or tracking problem, not of a ranking change. Real ranking changes are lumpy.
Slice by query group while you are in there. The free GSC regex generator turns a plain-English filter into the RE2 pattern Search Console will accept, which makes brand-versus-non-brand a two-minute check rather than a project.
Did it hit one surface or all of them?
Search, Discover and News are separate surfaces with separate eligibility rules, and the Performance report will show you each one on its own.
- Discover collapses, Search is flat. An eligibility or serving problem. Nothing about your rankings has changed.
- Search falls, Discover is flat. A ranking-side explanation is genuinely plausible. Keep going.
- Both fall on the same day by similar proportions. Look hard at your own site before you look at Google. Something upstream of both, such as a robots rule or a server problem, explains a synchronised drop better than an algorithm does.
Does a source Google does not compute agree?
This is the check people skip, and it is the one that would have saved a lot of wasted work in 2026.
Search Console is Google measuring Google. If a drop is real, it leaves fingerprints in systems Google has no hand in.
- Server logs. Fewer real sessions arriving from Google referrers is the hardest evidence available, and it is yours.
- GA4 or your analytics of choice. Organic sessions should fall in step. If they did not, the visitors never left.
- Revenue, signups, calls. The bluntest check there is. A traffic collapse that costs nothing usually did not happen.
When Search Console shows a cliff and every other system shows a normal week, you are looking at a reporting artefact. That is not a hypothetical failure mode, as 2026 has already demonstrated at scale.
The precedent: 50 weeks of impressions that were never real
On 3 April 2026 Google disclosed that a logging error had been preventing Search Console from accurately reporting impressions from 13 May 2025 onward. That is roughly 50 weeks of inflated numbers.
Google was specific about the blast radius, saying that clicks and other metrics were not affected and that the issue affected data logging only. The fix would roll out over the following weeks.
Follow the arithmetic, because it is worse than it first sounds.
- Impressions were too high, clicks were correct. So every CTR computed over that window, by anyone, was lower than the real figure.
- The correction looked exactly like a loss. Impressions in the Performance report visibly fell as the fix rolled out, on sites where nothing had changed.
- Any report built on that window is still wrong. Year-on-year comparisons that straddle it are comparing a real number against an inflated one.
I keep coming back to how many CTR optimisation projects ran during those 50 weeks, chasing a rate that was never as bad as the dashboard said. Nobody got an alert. The number simply looked disappointing, and teams did what teams do.
That is why I now treat "confirm it in a system Google does not own" as the first step rather than a due-diligence footnote. It costs ten minutes, and the failure mode it catches is a quarter of misdirected work.
Why Discover is the worst place to draw a conclusion
A lot of the loudest August reports concern Discover, which is the surface least suited to inferring anything about your site's health.
Discover traffic is volatile by design. There is no query behind it, no position to hold, and a feed decision Google makes per user. Its 2026 record makes the point.
- February 2026 Discover update. A confirmed update logged at 21 days 17 hours, covered in our February Discover update breakdown.
- March 2026 serving problem. Publishers saw sharp single-day drops in Discover impressions without having changed anything, and recovery was uneven across sites.
- No lever to pull. Discover exposes no rankings, no queries and no eligibility report, so there is nothing to diagnose in the way you would diagnose Search.
The practical rule: never let a Discover chart drive a site-wide conclusion. If Search is stable and Discover is not, you have a distribution problem on one surface, not a quality problem across your site.
What to do this week
Assume for a moment that some of the movement was real. The work that pays off is the same work that pays off during a confirmed core update, which is convenient, because you can start it before knowing.
- Timestamp the break precisely. Get the exact day, per surface, and write it down. A drop that starts on 30 July is not this event.
- Diff it against your own deploy log. Releases, migrations, plugin and theme updates, CDN and robots changes. This rules in or out the one cause you fully control.
- Confirm the loss outside Google. Logs, analytics, revenue. If they disagree with Search Console, stop and investigate the reporting.
- Segment the losers. Group lost pages by template, directory, topic and intent. A shared pattern is a lead; scattered losses across unrelated pages usually are not.
- Crawl the pages that lost. Check indexability, canonicals, redirect hops, internal links and Core Web Vitals on the losers specifically, and compare them against pages that held.
- Wait a week before concluding. Unconfirmed volatility frequently partially reverses, and the recovery is information too.
Step five is where most investigations stall, because Search Console knows which pages lost and nothing about why. Whether a losing page is orphaned, three redirect hops deep or quietly noindexed by a template change is a crawl-side question. The free page indexability checker answers it for one URL; a crawl answers it for all of them.
Ruling out a reporting artefact or your own last deploy needs Search Console, analytics and a crawl in one view. Kept in three tabs, the comparison never gets made.
CrawlRaven joins your Search Console performance, your GA4 traffic and a 200-point crawl into one ranked plan, so "these 40 pages lost" comes back with what is actually wrong with those 40 pages. Start free: 1 site, no credit card.
What not to do
Unconfirmed volatility has a reliable second act, which is a round of expensive changes made under pressure and never evaluated.
- Do not roll back an unrelated release. Removing a change that was not implicated adds a variable instead of removing one.
- Do not mass-edit content in week one. If you rewrite 200 pages now, you will never learn whether the drop reversed on its own.
- Do not deindex anything to "prune". Pruning is a considered strategy with a measurement plan, not a panic response to a five-day chart.
- Do not trust a single tracker. Volatility indices disagree constantly because they watch different keyword sets. One spiking index is noise.
- Do not file it as a penalty. Manual actions appear in Search Console with a notice. No notice means no manual action, whatever the drop looks like.
Key Takeaways
- →What is established: Third-party trackers moved sharply from 1 August 2026, peaked around 3 August and flared again on 5 and 6 August, with spikes across AccuRanker, Algoroo, AWR, CognitiveSEO, DataForSEO, Mangools, Mozcast, Semrush and Serpstat.
- →What Google says: Nothing. The Search Status Dashboard records no ranking update and no indexing, crawling or serving incident between 1 and 6 August 2026. The last confirmed ranking change is the June 2026 spam update, 24 to 26 June, logged at 2 days 1 hour.
- →Why one cause is unlikely: The reports span Search rankings, Discover, indexing and analytics. Those are four separate systems, and a single ranking update does not break all four in the same week.
- →The diagnostic: Check whether impressions or only clicks moved, whether one surface or all of them fell, and whether a system Google does not compute agrees that traffic was lost. Three of the four possible causes fall out of those checks.
- →The precedent: Google disclosed on 3 April 2026 that a logging error had inflated Search Console impressions from 13 May 2025 onward. Clicks were unaffected, so reported CTR was understated for roughly 50 weeks and the eventual impressions fall was a correction.
- →Do this now: Timestamp the break per surface, diff it against your deploy log, confirm the loss outside Google, segment the losing pages by template and topic, then crawl them. Wait a week before concluding anything about rankings.
Primary sources used in this post
- Google Search Status Dashboard: confirmed ranking updates and incidents
- Search Engine Land: Google June 2026 spam update done rolling out
- Search Engine Land: Google is fixing a Search Console bug that inflated impression counts
- Search Engine Roundtable: Google search ranking volatility, 1 August to 3 August
- Search Engine Watch: Google algorithm update in August 2026? What we know
Related reading on CrawlRaven
Frequently asked questions
Was there a Google algorithm update in August 2026?
Google has not confirmed one. The Google Search Status Dashboard records no ranking update and no indexing, crawling or serving incident for any day between 1 and 6 August 2026, and the last confirmed ranking change remains the June 2026 spam update, which ran from 24 June to 26 June and was logged at 2 days 1 hour. Third-party trackers did move sharply from 1 August, peaking around 3 August and flaring again on 5 and 6 August, but tracker movement is evidence that results changed, not evidence of an announced update.
If Google has not confirmed an update, why are the rank trackers spiking?
Rank trackers measure the results page, and the results page changes for many reasons that are not ranking updates. Google runs continuous smaller changes it never announces, tests layout and AI features, and reshuffles which surfaces appear for a query. A tracker cannot distinguish any of that from a core update, so a spike tells you the SERP moved and nothing about why.
How do I tell an algorithm hit from a Search Console bug?
Check average position and check a second data source. An algorithm hit moves your position, so impressions and clicks fall together and the losses cluster on pages sharing a template, topic or quality problem. A reporting bug leaves position and real-world behaviour untouched, so server logs, GA4 sessions and revenue keep pace while a Search Console metric steps to a new level. If the money did not move, treat the chart as suspect before you treat the site as broken.
What was the Search Console impressions bug?
On 3 April 2026 Google said a logging error was preventing Search Console from accurately reporting impressions from 13 May 2025 onward, and that the fix would roll out over the following weeks. Google stated that clicks and other metrics were not affected and that the issue affected data logging only. Because impressions were inflated and clicks were correct, every CTR calculated over that window came out lower than reality, and the visible fall in impressions when the fix landed was a correction rather than lost visibility.
My Google Discover traffic collapsed. Is that the same event?
Not necessarily, and Discover is the surface where that assumption is most expensive. Discover has its own confirmed history in 2026: a February 2026 Discover update logged at 21 days 17 hours, and a serving problem in March that produced sharp single-day drops for publishers who had changed nothing. A Discover collapse with unchanged average position in Search is evidence of an eligibility or serving problem, not of a ranking penalty.
Should I roll back changes I made in early August?
Only if the timing genuinely implicates them. Compare the break date against your deploy log first: a drop that starts the day after a release, hits one template or URL pattern, and coincides with a change to robots rules, canonicals, noindex directives or redirects is your change, not Google's. Rolling back an unrelated release during unconfirmed volatility adds a second variable to a situation you have not diagnosed yet.
How long should I wait before acting on unconfirmed volatility?
Wait for the data to stabilise, not for an announcement that may never come. Give it a week to two weeks of post-drop data before concluding anything about rankings, because unconfirmed volatility frequently partially reverses. Diagnosis is different from action: run the cheap checks immediately, because ruling out a reporting artefact or your own deploy costs nothing and can end the investigation on day one.
Does the Search Status Dashboard list everything Google changes?
No, and this is the limit worth understanding. The dashboard lists updates Google chooses to announce plus incidents it acknowledges, which is a small fraction of the changes shipped to Search in any given month. An empty dashboard means nothing was announced, not that nothing changed. It is still the only authoritative record available, so it belongs at the start of a diagnosis as a way to rule things in, never as proof that your site is imagining the drop.
15+ years of growing SaaS websites through SEO | Author, 200-Point Audit Checklist
Aditi has spent 15+ years helping SaaS companies scale organic traffic through technical SEO and content strategy. She is the author of the CrawlRaven 200-Point Audit checklist used by agencies and in-house teams to systematically improve search performance.