Google just spent two weeks rolling out a single update. So if your website traffic dropped suddenly inside that window, the numbers in your dashboard were never describing your store. They were describing a rollout still in progress. The hardest skill in SEO is sitting still while that happens.
The instinct is to fix something immediately. So you open your product pages and start rewriting titles. That instinct is understandable. However, it is also the single most expensive mistake you can make during a Google update.
Here is the uncomfortable part. During a rollout, a drop is not yet a verdict. It is a reading taken halfway through a process that has not finished running. Therefore the only honest answer to “what did I lose?” is that nobody knows yet, including Google watchers who do this full time.
This post explains what just happened, why the early numbers mislead, and the specific three-step hold that protects you. Moreover, it tells you the date when the data finally becomes trustworthy.
Before you change anything, get a baseline. Plug In SEO shows your Search Console rankings next to your on-page issues, so you can document today’s position in one screen.
- What happened: Google’s September 2026 spam update ran for roughly two weeks, from 24 September to 8 October 2026.
- Why your data lied: it landed in phases, not all at once, so each week told a different story.
- What to do: freeze changes, document everything, and judge nothing until 10 to 14 days after the rollout ends.
What actually happened to your traffic over the past two weeks
Google announced the September 2026 spam update on 24 September 2026. Then it confirmed the rollout was complete on 8 October 2026. That is close to a full fortnight of moving rankings.
The duration is the story. Google’s three previous spam updates this year each finished in about two days. This one took roughly two weeks, and several long-time trackers described its impact as closer to a core update than a typical spam update.
— Google Search Status Dashboard and Search Engine Roundtable
It also did not arrive evenly. Trackers logged distinct waves of movement around 25 to 27 September, again on 30 September, and again from 4 to 6 October. So a store owner checking on 28 September saw one picture. The same store on 7 October saw a different one.
Why the first week of traffic data lies to you
Two separate things distort what you see. Neither has anything to do with the quality of your store.
Updates land in phases, not in a single moment
A spam update does not flip one switch. Instead, Google evaluates different signals through systems that refresh on their own schedules. Consequently your rankings can fall in week one and partly recover in week two, purely because a later phase reached your pages.
That means a mid-rollout screenshot captures a half-finished state. Acting on it is like judging a recipe by tasting the batter.
Weekends change who is searching
The second distortion is simpler and widely underrated. Search behaviour shifts at weekends: mobile share rises, work-related queries fall away, and local results take up more of the screen. Therefore a Saturday-to-Sunday comparison against a Tuesday baseline will show a drop even in a week with no update at all.
The September rollout’s sharpest movement fell across a weekend. In other words, the scariest numbers many merchants saw were partly a calendar artefact sitting on top of a real update.
The three-step hold when traffic drops suddenly
None of this means do nothing forever. It means do these three things, in this order, and nothing else.
1. Freeze your changes
Stop editing while the ground is moving. Specifically, pause bulk rewrites of titles and meta descriptions, new redirects, URL handle changes, and mass content edits. These are exactly the changes that are hardest to unpick later.
The reason is diagnostic, not superstitious. If you change twenty things during a rollout and traffic recovers, you will never know whether you helped, whether you hurt, or whether the update simply finished. So you lose the ability to learn anything from the episode.
2. Document what your traffic is doing
Instead of editing, record. Take dated snapshots in Google Search Console covering impressions, clicks, average position and your top pages. Additionally, note which types of page moved, because that pattern is the actual diagnostic signal.
Collection pages behaving differently from product pages tells you something real. A single blended traffic number tells you almost nothing. This is also where having your rankings beside your on-page issues saves you a tab-switching mess later.
3. Set the date you are allowed to judge it
Write down a specific date before you look again. The convention among people who track updates closely is the rollout end plus 10 to 14 days of settling. For this update, that puts a fair verdict somewhere around 18 to 22 October 2026.
Put the date in your calendar. Then genuinely leave it alone until then, because a decision made on day three is a guess wearing a spreadsheet.
When a traffic drop is finally a real loss
After the settling window closes, the picture is worth trusting. At that point, compare against the snapshots you took rather than against memory.
| What you see after settling | What it most likely means |
|---|---|
| Traffic back near its old level | You saw rollout volatility. No action needed. |
| One page type down, others flat | A content or template issue on that page type. Start there. |
| Broad decline across the whole store | A site-wide quality or technical signal. Audit properly before editing. |
| Decline that began before 24 September | Not this update. Look for a change you made earlier. |
That last row matters more than merchants expect. Updates make convenient scapegoats for drops that actually started weeks earlier. A dated baseline is what separates the two.
What not to conclude from this
Waiting is not the same as ignoring. Some things still deserve action during any rollout, because they are broken regardless of what Google is doing.
- Genuinely broken things stay urgent. Pages returning errors, broken links and missing titles are defects, not update symptoms. Fix them.
- A freeze is not a content ban. Publishing a new blog post is fine. Rewriting two hundred existing product pages is not.
- Do not uninstall apps in a panic. Removing structured data or redirects mid-rollout adds a second variable to a measurement you already cannot read.
In short, fix what is objectively broken and leave what is merely scary.
Where to go next
Once the settling window closes and you want a structured look, our guide to running a Shopify SEO audit covers what to check and in what order. For the underlying work that survives any update, Shopify SEO best practices ranks the fixes by real impact. If you would rather start small, the 10 quick wins are safe to do at any time.
Frequently asked questions
My website traffic dropped suddenly. Was it the Google update?
Check the start date first. Google’s September 2026 spam update began on 24 September 2026, so a decline that clearly started before then is probably something else, such as a change you made or normal seasonality. If the drop lines up with the rollout window, treat it as unconfirmed until 10 to 14 days after the update finished.
How long should I wait before changing anything?
Until the rollout is confirmed complete, plus roughly 10 to 14 days of settling. This update completed on 8 October 2026, which puts a fair reading around 18 to 22 October. Waiting is not passivity, because you should be documenting the whole time.
Can I still publish new content during a rollout?
Yes. Publishing a new post or adding a new product is normal business and carries little risk. The thing to avoid is mass editing of pages that already rank, since bulk rewrites and new redirects make the eventual diagnosis much harder.
What is the difference between a spam update and a core update?
A core update is a broad change to how Google assesses content quality overall. A spam update targets content and tactics that breach Google’s spam policies. The practical difference for a merchant is small: both can move your rankings, and neither should be judged mid-rollout.
What should I record while I wait?
Dated Search Console snapshots of impressions, clicks, average position and top pages, plus a note of which page types moved. Record your current titles and meta descriptions too. Then you have a real before-and-after rather than an argument about what the site used to look like.
Waiting works better with a record. Plug In SEO puts your Search Console rankings next to your on-page issues, so your baseline and your fix list live in the same place.