SEO Trends for 2026: What Your Business Needs to Know

Most businesses are doing the same SEO work they were doing two years ago and getting fewer visitors for it. That is not a failure of execution. Between 2024 and 2026 the mechanics of search changed underneath everyone at once, and the pages that used to earn a click now often answer the question before the click happens. Ahrefs measured a 58 percent drop in click-through rate for the top-ranking result when an AI Overview appears above it.

What follows is not a list of predictions. Every shift below already happened, has been measured by more than one independent study, and can be checked against your own analytics this afternoon. If you are still deciding how the different channels fit together, our breakdown of the difference between SEM, SEO and PPC is the better starting point. If you already run SEO and the numbers stopped making sense, keep reading. At CIOR Digital, a digital growth agency in La Jolla, this is the conversation we have with almost every client who walks in with a traffic chart that bends the wrong way.

Before anything else, there is one question worth answering honestly: has your ranking actually fallen, or are you ranking exactly where you always did and simply receiving fewer visitors for it? Open Search Console, compare impressions against clicks over the last eighteen months, and look at which line moved. If impressions held steady or grew while clicks fell away, you are not looking at a ranking problem and no amount of conventional optimisation will fix it. That single comparison determines which half of this article is relevant to you.

What actually changed in search

The short version: the search results page started answering questions itself, and the click became optional. Google now places an AI-generated summary above the organic results on more than 20 percent of searches, according to SparkToro’s analysis of US query behaviour in early 2026. On those searches, the summary frequently satisfies the user completely.

The knock-on effect is measurable at the level of the whole market. SparkToro puts zero-click searches, meaning searches that end without any click at all, at 68 percent of US queries in early 2026, up from just over 60 percent two years earlier. Those eight percentage points represent an enormous volume of visits that simply stopped existing.

Running alongside the summaries is a second surface that behaves quite differently. AI Mode replaces the results page entirely with a conversational interface, and Google stated at its 2026 developer conference that it had passed a billion monthly users with query volume more than doubling each quarter. That figure comes from Google itself and should be read accordingly, but the direction is corroborated by independent measurement. What matters commercially is the structural difference: in AI Mode there are no organic listings sitting below the answer, so a brand that is not named inside the response has no presence in that conversation at all.

None of this means the underlying work changed. Google still needs to crawl your site, still evaluates the same quality signals, and still ranks pages. What changed is what happens after the ranking. A page that ranks first and is not referenced in the summary above it is now competing with an answer the user has already read.

AI Overviews decide who gets the click

When an AI Overview appears, the click-through rate for the pages beneath it falls sharply, and the size of that fall is the most argued-about number in the industry. Ahrefs, working from 300,000 keywords with data collected in December 2025, put the drop for the top-ranking result at 58 percent. The same team measured 34.5 percent in April 2025, which tells you the trend has been moving in one direction as the feature rolled out to more countries and languages.

Independent confirmation matters here, because SEO vendors have an obvious interest in the story. The Pew Research Center, which sells nothing, tracked real browsing behaviour in July 2025 and found that users clicked a result in 8 percent of visits when an AI Overview was present, against 15 percent when it was not. Different method, different sample, same direction.

There is a second half to this story that most coverage skipped. Seer Interactive tracked 53 brands across 5.47 million queries and 2.43 billion impressions from January 2025 to February 2026, and found that click-through rate on AI Overview queries bottomed out at 1.3 percent in December 2025 before recovering to 2.4 percent by February 2026. The floor appears to have been reached. That is not a return to normal, since searches with no AI Overview still convert attention into clicks at a noticeably higher rate, but it does mean the decline is not infinite.

It is worth understanding why these figures disagree, because the disagreement is not evidence that someone is wrong. Each study draws on a different panel, measures a different mix of query types, and covers a different window during a rollout that is still expanding across countries and languages. Ahrefs works from aggregated Search Console data on informational keywords. Pew observes real browsing sessions. Seer tracks a fixed set of brands over time. Any business quoting one of these numbers as its own expected loss is misreading them. The useful takeaway is the range and its direction, not a single figure to plug into a forecast.

Sources: Ahrefs AI Overviews study, Pew Research Center analysis, SparkToro zero-click research.

Citation replaced ranking as the goal

The most useful number in the Seer data is not the decline, it is the gap between being cited and being ignored. On a results page carrying an AI Overview, a page that the summary cites earns roughly 2.1 percent click-through rate. A page sitting on that same results page without being cited earns roughly 0.9 percent. Same page, same position, more than double the clicks, decided entirely by whether the summary referenced it.

That single comparison reframes the objective. Ranking has not stopped mattering, because AI systems overwhelmingly draw their citations from pages that already rank well, so a page outside the top twenty is rarely in the running at all. Ranking is now the qualifying round rather than the prize. The prize is being the source the answer quotes.

This is what generative engine optimization means in practice, and it is a genuinely different discipline from keyword optimisation. It concerns how a page is structured, whether it answers a specific question in a self-contained way, whether it carries facts a model can attribute, and whether your brand is described consistently enough across the web for a model to treat it as an entity it knows. Building an AI-era search strategy starts with accepting that those four things are now the work, rather than treating them as refinements on top of it.

There is also a category of visibility here that produces no click at all and still produces business. When an assistant recommends your company by name without linking to you, the user often searches for you afterwards, arrives through a branded search, and gets recorded as organic traffic that appears to have come from nowhere. That mention is a real commercial event and it is invisible in every analytics tool you own. It is the main reason that judging this shift purely by referral sessions will understate what is happening to your brand.

• Rank firstCitation is drawn from pages that already perform. Without top-twenty visibility there is nothing to cite.
• Answer in a quotable formA self-contained answer in the opening lines of a section can be lifted. A buried answer cannot.
• Carry attributable factsFigures with a named source and a date are what models reach for. Unsourced assertions are not.
• Be a consistent entityYour business needs to be described the same way across your site, your profiles and third party mentions.

Less traffic, better traffic

Here is the part that surprises people: the visits that survive the AI answer tend to be worth considerably more than the ones that disappeared. Multiple independent studies through 2025 and 2026 found visitors arriving from AI platforms converting at several times the rate of standard organic traffic. Semrush put the average at around 4.4 times. Adobe Analytics, working across more than a trillion visits to US retail sites, found AI-referred shoppers converting 42 percent better than non-AI traffic in March 2026, having found them converting 38 percent worse exactly twelve months earlier.

The honest version of this finding includes its caveats, and there are several. The range across studies runs from roughly 1.3 times to more than 20 times, which is far too wide to treat as a single benchmark. The studies measure different conversion events, from newsletter sign-ups to completed purchases. Most importantly, the volumes are small: on a typical business site, AI referral traffic is currently around 1 percent of total sessions, so small absolute numbers produce volatile percentages.

The mechanism behind the effect is straightforward once you picture it. Someone who asks an AI assistant which agency handles technical SEO for medical clinics, reads a synthesised comparison, and then clicks through to you has already done their shortlisting. A person who types the same query into Google and clicks your result is at the beginning of a research process that will run across several tabs and several days. The AI visitor arrives further down the same path, which is why they behave differently when they land.

The practical implication is a change in how you judge a bad month. A twenty percent fall in organic sessions used to be an unambiguous problem. Today it can accompany flat or rising revenue, because the visits that vanished were largely the ones that were never going to buy anything. Reporting that leads with session counts will keep generating alarm about a situation that is commercially neutral, or occasionally an improvement. Reporting that leads with conversions and revenue per session tells you what actually happened.

Technical work is now the entry ticket

Technical SEO used to be a way to gain an edge. It is now the condition for being in the conversation. A page that a model cannot crawl, cannot parse, or cannot render without executing JavaScript is a page that will not be cited, regardless of how good the content is. The bar has not risen dramatically, but the penalty for sitting below it has.

Page speed belongs in this category too, and it needs reframing. Core Web Vitals are a modest ranking signal and always were. What they genuinely determine is whether the smaller, more valuable stream of visitors you now receive actually stays long enough to convert. When every visit is scarcer, losing a share of them to a slow first load costs more than it did when traffic was plentiful.

One correction belongs here, because the previous edition of this article got it wrong and plenty of marketing advice still repeats the error. Third party cookies were not phased out. Google abandoned the deprecation plan in July 2024, confirmed in April 2025 that it would not introduce a choice prompt in Chrome, and shut down most of the Privacy Sandbox APIs it had built as replacements in October 2025. Cookies still work in Chrome by default. The first party data argument survives anyway, since Safari and Firefox have blocked third party cookies for years and consent requirements apply regardless of what any browser does, but the deadline everyone was planning around never arrived.

• Server-side rendering for content that matters. If the text only appears after JavaScript runs, assume some systems will never see it.
• Crawlability for AI agents. Check your robots.txt and any bot-blocking rules at CDN level. Plenty of sites are blocking the crawlers they want to be read by.
• Structured data that describes reality. Article, Organization, LocalBusiness, FAQ and Product schema give machines an unambiguous description of what a page is.
• Clean heading hierarchy. One H1, logical H2s that read as answerable questions, no decorative heading levels.
• Fast first render. Not for the ranking points. For the visitor you worked much harder to earn.

Entity consistency feeds the AI answer

For any business that serves a geographic area, the biggest available advantage right now is boring and unglamorous: describe yourself identically everywhere. AI systems assemble their picture of your business from your website, your Google Business Profile, directories, review platforms and press mentions. Inconsistency across those sources does not just weaken the picture, it can prevent a model from being confident enough to name you at all.

Voice search, which absorbed years of industry attention, has quietly resolved into this. Conversational queries did become normal, but they became normal inside AI assistants rather than through voice-specific optimisation. The practical work is the same work: answer real questions in natural language, and keep your business details consistent so an assistant can retrieve them.

Businesses without a physical service area are not exempt from this, they simply have a harder version of the same task. Without a Business Profile anchoring the entity, the description a model assembles comes from your own site, your LinkedIn presence, industry directories and whatever third parties have written about you. Where those sources contradict each other on what you do or who you serve, the model has no way to resolve the conflict and tends to reach for a competitor it can describe with more confidence.

• Google Business Profile, complete and current. Categories, hours, services, and photos that are not five years old.
• Identical name, address and phone number everywhere. Including abbreviations and suite numbers.
• Reviews, and replies to reviews. They are read as evidence of an operating business, not just as social proof.
• A description of what you do that does not change between platforms. Same positioning, same words, same services.

Your dashboard is lying to you

Most analytics setups are currently misreporting the channel that is growing fastest. Traffic arriving from AI assistants frequently lands in GA4 as direct or as generic referral, because there is no default channel grouping for it. The result is that organic looks slightly worse than it is, direct looks inexplicably healthy, and the AI channel appears not to exist. Surveys through 2025 and 2026 consistently found that only around 14 to 16 percent of brands were tracking AI search as a distinct channel.

Fixing this takes an afternoon and changes the conversation, because you cannot argue for investment in something you are not measuring. Once the channel is isolated, the comparison to make is not volume against organic, which AI traffic will lose badly, but conversion rate and revenue per session, which it often wins. That is the same reasoning that applies to measuring the return on SEO properly rather than by traffic alone.

There is a time cost to delaying this that is easy to miss. Analytics platforms do not backfill a channel definition you create today, so the months you spend without one are months you will never be able to analyse properly. A business that sets this up now will have a year of comparable data when the question of how much to invest in AI visibility becomes unavoidable. A business that sets it up next year will be making that decision on a few weeks of numbers.

• Create a custom channel group in GA4Classify sessions from chatgpt.com, perplexity.ai, gemini.google.com, claude.ai and copilot as a distinct AI Referral channel.
• Compare quality, not volumeLook at conversion rate, pages per session and revenue per session against organic, not raw sessions.
• Add a discovery question at the point of conversionAsk new leads how they found you. It catches the influence that never shows up as a click at all.
• Check your citation rate separatelyBeing mentioned without being linked is still a business outcome, and it will not appear in any analytics tool.

What to do first, depending on your business

The correct response to all of this depends almost entirely on your size and your market, and the wrong response is to do everything at once. A single-location service business and a multi-service company competing nationally have different first moves, and only one of them should be thinking about generative engine optimization this quarter.

The common thread is sequence. Measurement comes before strategy, because without an isolated view of AI traffic you are making decisions on a distorted picture. Foundations come before ambition, because citation is drawn from pages that already rank and load. Everything else follows from those two. If you would rather have that sequence mapped against your actual situation than guess at it, that is the sort of thing a growth partner for service businesses should be doing with you before any work starts.

What did not change

It would be easy to read all of this as a reason to abandon what you are doing. That would be a costly misreading. Every mechanism described above still selects its sources from pages that rank, load properly and answer the question well. The fundamentals did not get replaced, they got promoted into prerequisites.

Be equally sceptical of the opposite claim. Anyone telling you that SEO is finished is usually selling the replacement, and anyone telling you nothing has changed has not looked at their click-through rates recently. The measured position sits between those two, and it is not complicated: the same work, judged by different outcomes.