A pattern has come up repeatedly with clients this year. Traffic is down twenty or thirty percent. They open their rank tracker, and the rankings look fine. Nothing obviously moved. The site was not hacked, nothing was deindexed, and no one changed anything.
That combination is new, and it has a specific cause. Understanding which of two things happened to you decides whether you have work to do on your content or not.
What actually happened in 2026
Two forces hit at once.
Google ran two broad core updates. The first rolled out from 27 March to 8 April, took twelve days, and produced some of the highest ranking volatility the tracking tools have recorded. The second started on 21 May. Both moved a lot of results, and small and medium business sites in competitive niches were among the more affected groups.
The results page itself changed. The May update arrived in the same week that AI Mode became the default search experience and the search box was rebuilt for the first time in twenty five years. AI Overviews now answer a growing share of queries directly on the page.
That second change is the one people misdiagnose. It does not move your ranking. It removes the click that used to come with your ranking.
The fifteen minute diagnostic
You do not need a paid tool for this. Google Search Console has what you need.
- Open Performance, then Search Results.
- Turn on all four metrics: clicks, impressions, average CTR, average position.
- Set a comparison of the four weeks after the drop against the four weeks before.
- Read them together rather than one at a time.
Then match your pattern:
- Position fell and clicks fell. Core update. Your pages are being ranked lower than they were, and the fix is content quality.
- Position held, impressions held, clicks fell, CTR fell. The results page changed. You are still being shown, people are not clicking through, because they got the answer above your link.
- Impressions fell but position held. Fewer people are searching, or Google is showing your page for fewer query variations. Check seasonality before panicking.
- Everything fell off a cliff on one day. That is usually technical. A robots file, a noindex tag, an expired certificate, a migration nobody told you about.
If it was a core update
The pattern in what survived is consistent, and it is not subtle. Sites that held or gained had original information, first hand experience, clear authorship, real sourcing, and content aimed at a genuine need. Sites that fell tended to be publishing volume that restated what was already on the web.
Practical steps, in order of impact:
- Audit content honestly. For each underperforming page, ask what it contains that could not be found on ten other sites. If the answer is nothing, improve it or remove it. Thin pages drag the whole domain.
- Add what only you have. Your own numbers, your own client results, photographs of your actual work, prices you actually charge. We publish real price ranges for exactly this reason.
- Make the author real. A named person with a genuine bio and a track record, not a generic team byline.
- Merge overlapping pages. Three thin articles on the same topic compete with each other. One strong page beats them all.
If it was AI answers
This one is harder because there is no ranking to recover. The strategy changes shape instead.
Look at which pages lost clicks. There is usually a clean split. Pages that answer a simple factual question get absorbed into AI answers. Pages where the reader has to see something, compare something, or contact someone hold up. A page explaining what a term means will be summarised. A page showing your work and your pricing will not be, because that is not a summarisable answer, it is a decision.
So the response is a rebalance rather than a rescue:
- Stop investing in pages whose only value was answering a question in one paragraph.
- Invest in pages that require the reader to act: comparisons, case studies, pricing, process, proof.
- Structure content so it can be quoted accurately when it is used as a source. Clear headings, direct answers, specifics rather than hedging.
- Track brand searches and direct traffic, not only organic clicks. Being cited in an answer creates demand that shows up under a different label.
The uncomfortable part
Analyst forecasts put a meaningful share of search traffic shifting toward AI assistants by the end of this year. Whether the exact figure holds is beside the point. The direction is not in question, and a business that treats organic clicks as the only measure of search value is measuring a shrinking thing.
The reassuring part is that the work does not split into two jobs. What earns a citation in an AI answer is what survives a core update: specific, sourced, first hand, trustworthy content from a business that is clearly a real entity. There is no separate trick. The bar just moved up.
Frequently asked questions
How do I tell if a core update hit my site?
Open Google Search Console, go to Performance and then Search Results, and compare the two weeks after the update against the two weeks before. Look at clicks, impressions and average position together. If average position dropped alongside clicks, that points to a core update. If average position held and only clicks fell, the result page changed rather than your ranking.
When were the 2026 Google core updates?
The March 2026 core update rolled out from 27 March to 8 April, taking twelve days, and was the first broad core update of the year. A second core update began on 21 May and completed in early June. A spam update also ran in March.
Why did my clicks drop when my ranking stayed the same?
Because the search results page changed around you. AI Overviews and AI Mode answer more queries directly, so a searcher gets what they needed before clicking anything. Your position is unchanged, the traffic behind that position is not. Informational pages that answer one simple question are hit hardest.
Can I recover traffic lost to a core update?
Sometimes, but not by making technical tweaks. Core update recovery comes from genuinely improving the content: original information, first hand experience, clear authorship, real evidence. Thin pages that restate what is already on the web rarely come back and are often better removed or merged.
Is SEO still worth doing in 2026?
Yes, but the goal has widened. You are now optimising to be the source an answer is built from, not only the link someone clicks. The qualities that earn that are the same ones that survive core updates, which means there is one job here rather than two.