1. SEJ
  2.  ⋅ 
  3. SEO

Date: Confirmed, Rating: Meets Expectations, Comments: None

Google's core updates arrive with a start time and a scope. Everything else you've read about them came from somewhere else.

Date: Confirmed, Rating: Meets Expectations, Comments: None

The routine is familiar by now. Google posts a sentence. It has a name, a start time, and a note that the rollout may take up to two weeks. Within the hour, somebody has published a “what we know so far.” By the evening, there’s a list of winners and losers. Before the week is out, with the rollout still running, there’s a recovery guide in seven steps.

I’ve watched this for most of 20 years. The only part that has changed is the sentence. It used to be longer.

The Announcement Used To Say Something

When Google launched what the industry later called Panda, in February 2011, they wrote a post about it. The post said what the change was for: sites that add little, copy from others, or just aren’t very useful. It said how big it was: 11.8% of queries. You could argue with the result, and plenty did, but there was a stated target to argue with.

Here is the entry for the November 2026 core update, which started on Nov. 10. It was released. It applies globally and in all languages. It may take up to two weeks. That’s the lot: three facts, none of them about your site.

Google has been open about why. In a Q&A on Search updates, it said it changes Search thousands of times a year and announces the ones it considers notable, about 10 a year when it last counted in public. Its core updates documentation adds that smaller core updates go out continually and aren’t announced at all. So a named update is a ranking change that Google decided to put on a calendar. The calendar entry tells you when. It tells you nothing about what changed, or about what happens in between.

The announcement shrank to a timestamp, and the commentary expanded to fill the space.

What The Date Is Good For

The date does have a use. Google’s own guidance uses it: wait a week after the rollout finishes, then compare that week with the week before it started. The date draws a line through your data so you know where to look. Visibility trackers do a similar job across the market; they show which sites moved.

What neither of them shows is why. Which system changed, and why it moved your site in particular, is known inside Google and nowhere else. A rank tracker and a strong feeling won’t close that gap.

That hasn’t slowed anyone down. An obscure tool vendor published a winners-and-losers study of the November update within a day or two of the launch. It gave the wrong start date and called it the sixth core update of the year. Google’s dashboard lists three. I’d say the analysis was rushed, but that implies there was something to analyze.

Review Day

The closest thing I know to a core update is the annual performance review.

Anyone who has had a job knows the meeting. You walk in and get a rating that was settled long before the meeting was booked. It reflects the year you had. It also reflects the people you were measured against, because plenty of companies calibrate ratings across a team; you can do the same work as last year and drop a grade because three colleagues did better. Search works the same way. A page is ranked against every other page that could answer the query, so a site can lose ground without changing anything.

A core update is review day. It’s when the rating becomes visible. Nobody sensible leaves that meeting and asks what changed about the meeting.

The comparison stops in one place. After a review you get a manager, written feedback and a chance to ask what went wrong. After a core update you get a dashboard entry. That missing conversation is the entire market for update commentary: an industry of people offering to be the manager who explains your rating, none of whom were in the calibration room.

The Damage Is Real

The strongest objection to all this is that the losses are real, and it’s correct. Companies do lose a large share of their organic revenue overnight, and people lose jobs over it. The practitioners who spend their time on recoveries are dealing with real businesses in real trouble.

There’s also evidence that sometimes the reviewer gets it wrong. In October 2024, Google invited a group of independent publishers hit by the September 2023 helpful content update to its campus. According to people who were in the room, they were told their content wasn’t the problem, and that they shouldn’t count on getting their traffic back. That is an awkward fact for anyone, me included, who says most affected sites had it coming.

Look at what it does to the recovery guide, though. If the content wasn’t the issue, “seven steps for improving your content” were never going to work. The publishers with the best case against Google are also the best evidence that nobody outside it could diagnose them.

Who Needs The Explanation

Now the less sympathetic group, which in my experience is the bigger one.

Most of the affected sites I’ve been asked to look at over 20 years had the same history. Somebody decided what they could get away with before finding out what the market already offered. They hadn’t researched what customers needed or looked at how good the existing answers were. They picked keywords with volume, wrote about them, and launched.

When that didn’t work, they went looking for a shortcut, and social media had plenty. A thread with a promise that was too good to be true, and was. A “hack” from somebody who discovered SEO in the summer and by autumn had a conference slot and a YouTube channel. In office terms, that’s career advice from the colleague who joined in August and already runs a workshop. They copied it without asking why it would work. Some of these tricks do work for a few months, which is the worst outcome available, because the fall then lands inside an update window and gets blamed on the update.

Then they hired SEO to make it work, as if rankings were a delivery service for businesses that forgot to have a product. I’ve written about the wall that approach runs into; it hasn’t moved.

This is the employee who did the minimum all year and arrives at the review certain that the rating system has changed, and has brought printouts. In fairness to them, they have a problem that needs solving. “The algorithm changed” is something you can say to a board. “We built a site with nothing to say and then paid someone to get it found” tends not to make the deck.

So there’s a buyer for an outside cause, and the update coverage supplies one. It comes with a diagnosis nobody outside Google could have made and a retainer that renews before the next update can contradict it. The seller gets a client, and the client gets a story. The site stays exactly as useful as it was, which was the problem.

Why The Explanation Looks Necessary

This trade feels legitimate to the people in it because of how the industry thinks about quality.

Take E-E-A-T. Google coined it for its quality raters, and Google’s own documentation says it isn’t a specific ranking factor. The industry read that and sold E-E-A-T audits anyway, with scores for a thing Google says it has no score for. Somewhere there is an author bio that was added to fix a business model.

The traits themselves are older than the initialism. Peter Morville’s user experience honeycomb had “credible,” “useful,” and “valuable” in it in 2004, as facets of a product, with no ranking attached. The industry needed Google to hand it an initialism before it noticed that people prefer things they can trust.

“JavaScript SEO” has the same shape: it’s progressive enhancement, which web developers were practicing before anyone gave it a conference track. I made that case in Stop Learning SEO and won’t repeat it.

If quality is a list of items to tick, then a core update is a new edition of the list, and you need someone to tell you what’s on it. There is never a shortage of volunteers. Back in the office, that’s working to the KPIs. The person who works to the KPIs has to know the moment they change. The person who is good at the job mostly finds out on review day that they still are.

So people who hold quality as a checklist need each update explained; the announcement explains nothing, and so somebody makes one up.

Two Questions

The first question a business could ask after an update is “What did this update change for my site?” Answering it would take the list of systems that changed, how much each one counts, and what every competing page did in the same fortnight. Nobody outside Google has those. You can still pay for an answer. It’ll arrive as a confident PDF with a traffic-light table, and it will tell you to improve content quality, show more expertise, and prune the thin pages.

Now read the PDF from the update before. It says the same thing. So will the one after the next update, because the advice was never derived from the update. If the remedy is the same whatever changed, the change taught nobody anything.

The second question is “what do the people I want as customers need that they can’t already get?” It’s a worse question to be asked. It has no villain and no deadline; it can’t be handed to a thread, and the answer may be that the business has nothing to add yet. It’s also slower and more expensive to answer than a PDF. But it can be answered by people you can find and talk to, and it could have been answered before the site launched.

The fair objection to the second question is Google. Google is in the business of taking in what other people know and serving it back as its own answer, and AI Overviews and AI Mode have made that the main event. The part of the web where people found you through a search result is being eaten, and Google is the one holding the fork. I’ve covered what that did to publishers’ traffic. So why polish anything for a channel that’s shrinking?

Because the second question was never about the channel. A site is a product, and what people need from it doesn’t change when fewer of them arrive through Google. What does change is how much of what you know you should hand over for free. I’ve argued that publishing everything is a terrible business strategy, and I still think so. Holding back the method is sensible. Shipping a site that’s a bare outline of what you’re supposed to be good at is a different decision, and it fails the customer long before it fails a ranking system.

The line between those two is thin, and it’s meant to be. It sits in a different place for every business, and no update, checklist, or thread will tell you where yours is. A search engine certainly won’t; it looks after its own interests, as it always has. The only people who know how much is enough are the ones you’re trying to sell to, which is one more reason to go and ask them.

In the office, it’s the gap between “what did the calibration committee change this year?” and “what did I do this year?” Only one of those is yours to find out, and it’s the one that doesn’t come with somebody else to blame.

Google will post another sentence soon enough. The recovery guide for it already exists. It’s the one from last time.

More Resources:


This post was originally published on The Inference.


Featured Image: Roman Samborskyi/Shutterstock

Category SEO
Pedro Dias Independent Consultant at Visively

I help companies design systems that make their content findable — by both search engines and AI (Artificial Intelligence). I’m ...