Marketing executives, set your pitchforks down for a second. Take a deep breath. Let your SEOs take a beat. Here’s the situation: right now, the SEO industry is in the middle of one of its biggest transformations ever, and it’s much more massive than just another core update.
For decades, we SEOs have relied on a pretty consistent set of KPIs: clicks, impressions, rankings, and traffic. There would be occasional dips or adjustments that needed to be made, but for the most part, SEO reporting and analysis were fairly straightforward. As if 2025 couldn’t get any worse, though, those metrics are no longer telling the full story.
Between AI search (and social search), changing user behavior, and “zero-click” search experiences, your SEO dashboards may be looking a little bit messier these days. Not to fearmonger, but there’s no going back; we’re past the point of no return. If you’re like me (a Taurus who doesn’t do well with change), you’re also realizing that it’s time to stop holding onto those precious KPIs we’re so used to and rethink how we do literally everything.
The Perfect Storm That’s Reshaping SEO
Before we dig into how we’re supposed to effectively turn everything we know upside down, it’s important to preface it with why we need to do that in the first place. If you’re an SEO reading this, I’m sure you know all of these things already; if you’re not, welcome to our world. It’s (usually) cozy here.
Changing User Behavior
I know it sounds vague to just say “users are behaving differently,” but that’s truly the best way to sum it up. They’re searching less (on traditional search engines), expecting more immediate answers, and interacting directly with AI chatbots or overviews rather than with websites.
How it looks for SEOs: If you’ve noticed clicks declining, you can likely attribute that to zero-click search. The new hurdle becomes figuring out where users are going and how you can meet them there.
AI as a Search Tool
Large Language Models (LLMs) like ChatGPT, Perplexity, and Gemini are being used as an alternate (and in some cases, preferred) search interface. Doubling down on the user behavior shift, these platforms are preferred because they decrease the mental workload for the user.
If users don’t need to figure out what to search, type it in, click around on the top ranking results, and (God forbid) read a few webpages to find what they’re looking for, they won’t do it. It’s human nature; it’s the path of least resistance.
How it looks for SEOs: You may have seen an increase in Referral traffic in your GA4 reports. Double-click into that, and you’ll get an idea of where your users are coming from. And if you know what’s good for you, go ahead and set up that LLM Search custom channel grouping now.
The Clickless Search Era
Zero-click search is happening, whether it’s because users don’t feel the need to click or there’s nowhere for them to click. Let me explain:
- Google’s SERPs are increasingly dominated by AI Overviews, featured snippets, and other widgets that satisfy user intent without a click.
- If a user is searching using an LLM, there’s a high chance that the answer provided doesn’t include citations. Other than Perplexity (which is designed to always cite sources for the information it provides), other LLMs like ChatGPT are pretty unlikely to cite sources without being directly asked to.
How this looks for SEOs: The traditional impression > click > engagement > conversion funnel has exploded. With nowhere (or no need) to click, the user journey becomes fragmented and unpredictable. A user might see your brand in an AI Overview, wait a week, and then go to your website directly, or look up your Instagram to find you. How in the world are we meant to track that, let alone attribute the conversion to a certain channel?
Google Ruining Our Lives (Again)
In a September 2025 update, Google has removed the “num=100” parameter from searches. That means that the top 100 results are no longer appearing on the SERP, just the top 10. The other 90? Given that 91.5% of people never even click to page 2 of Google, they may as well not exist.
For a platform that’s worshipped by and obsessed over by millions of people worldwide, Google really loves to hurt our feelings.
How it looks for SEOs: If your impressions tanked in September, you can blame our robot overlord. Not only is Google Search Console looking more and more like a tachycardic patient’s heart monitor these days, but now, tools like Semrush and Ahrefs aren’t able to track rankings down to the 100th position without using 10x the server power they were before.
This all results in a bunch of panicked SEOs, anxious to explain to their clients and their managers why performance looks so bleak (or is that just me?). The reality is, SEO performance isn’t declining; it just hasn’t been redefined quite yet.
So let’s do that.
The Devaluation of Traditional SEO Metrics
Let’s look at some of the metrics we’re used to, and talk about why they might not necessarily be the most reliable or paint the most accurate picture anymore.
1. Clicks: The First Domino to Fall
As we discussed above, zero-click searches are the new normal: AI overviews often resolve intent without a single page visit, and LLMs are questionable at best when it comes to citing their sources.
- The Result: Organic CTRs are dropping across industries.
- The Takeaway: Clicks still matter (after all, that’s how you definitively know that people are going to your site), but their quantity needs to be devalued as a performance metric in favor of quality.
2. Impressions: Losing Accuracy & Meaning
Between the removal of the num=100 parameter and the movement of search to AI and social platforms, impressions are quickly becoming less reliable as an SEO KPI. Tools like Ahrefs, Semrush, and other third-party crawlers can no longer fetch full SERP datasets.
- The Result: We no longer have a complete or consistent picture of true visibility, leaving SEOs with no way of knowing how often content is actually seen.
- The Takeaway: Impressions shouldn’t be abandoned entirely, but they must be interpreted cautiously. Until Google or third-party tools adapt, treat them as directional indicators, not end-all-be-all measures of performance.
3. Rankings: Vanity Metrics on AI-Dominated SERPs
Ranking first used to mean everything (cue the PTSD of being asked “why aren’t we first?” daily); it guaranteed visibility, authority, and traffic. Now, AI overviews, featured snippets, and other SERP features have rewritten the playbook.
Even if your content technically “ranks first,” you’re likely buried below the fold by Google’s own AI summaries. On top of that, that pesky num=100 change makes it nearly impossible to track rankings past position 10 with any real accuracy.
- The Result: The concept of “ranking” as we know it no longer correlates with actual visibility or user engagement.
- The Takeaway: Shift your focus from position to presence. Stop aiming for first-place rankings only for it not to drive any real traffic to your site (remember the zero-click thing?); instead, turn your strategy into one that focuses on being found across each surface where users search.
4. Traffic Attribution: The Invisible Traffic Problem
GA4 hasn’t quite caught up to the AI search landscape yet (I’m waiting patiently). There’s no dedicated channel grouping for AI traffic, meaning that visits from AI search platforms and LLMs often show up in the Referral, Direct, or Unattributed buckets.
- The Result: Marketers are underreporting their search presence and making strategic decisions based on incomplete data. If you’re noticing a rise in Referral, Direct, or Unattributed channels and not double-clicking to investigate further, 🚨you’re doing yourself a huge disservice🚨.
- The Takeaway: Until Google catches up with… well, itself (it doesn’t make sense to me either), SEOs need to get proactive: customize channel groupings, audit referral traffic, and re-bucket visits that originate from AI search engines. Slack your data analysts and get to work; attribution hygiene is now part of SEO.
Reinterpreting SEO Measurement: What the Hell Do We Do Now?
Not to play the victim card, but between getting leadership buy-in concerning AI search, having to explain sudden shifts in performance without sounding like a broken record, and restructuring basically the entire way you’ve been doing everything up to this point, being an SEO is tough right now.
Fear not; myself and the NoGood SEO team have done some serious noodling over it, and I think we’ve come up with a pretty good plan for how we can slow the pace with which the world is burning. It goes something like this:
Step 1: Don’t Panic Over Benchmarks, Redefine Them
It’s time to let go. Accept that traditional metrics that are already experiencing drop-offs will likely continue to trend lower. This is a new baseline, not a failure.
Slap on that pattern recognition hat and focus on understanding directionality, not absolute numbers. I know that this fluidity can take a while to get used to, especially for my particularly number-brained SEOs out there (present), but it’s a necessary step to understanding your brand’s overall presence in the digital landscape.
Here are some examples:
- If a page on your site is seeing fewer and fewer clicks year-over-year, dig deeper to look at engagement and conversion; significant increases in either area could indicate that while search visibility declined, traffic quality improved.
- A low CTR doesn’t sound great as a metric, but when paired with higher average engagement time and scroll depth, that’s a signal of content resonance.
- Branded keyword rankings and traffic have been largely ignored over the years; now, however, an increase in branded traffic could signal that users are seeing your brand name (uncited) in an LLM and moving to traditional search to look you up directly.
- If impressions remain steady while clicks fall, it could suggest zero-click behavior rather than declining visibility. Use that as your guide for digging deeper to see if that’s the case or not.
Step 2: Wait & Watch as Tools Evolve
Right now, it’s simply too early to know how analytics and SEO platforms will adapt to the num=100 change. Whether it’s via Reddit threads, our LinkedIn, or SEO communities you’re a part of, keep an eye on how Ahrefs, Semrush, and Google Search Console evolve in the coming months.
Step 3: Shift Focus From Rankings to Brand Visibility
Shake off that keyword rank obsession and opt for holistic visibility analysis across earned, owned, and AI surfaces. Track where and how your brand shows up in AI Overviews, answer engines, Featured Snippets, and other digital ecosystems.
I know that explanation sounds vague, but here are a few tips to do this:
- Use an AI visibility monitoring tool like Goodie (yes, we’re shamelessly promoting stuff in 2025) to track mentions in AI overviews or featured answers, even if they don’t link back; these are signs of topical authority.
- Social listening is key; track brand mentions in Reddit discussions or Quora threads that appear in SERPs. Not only are these social and UGC sites among the top cited domains in LLMs, they’re also a form of secondary organic visibility.
Step 4: Clean Up That GA4 With Custom Channel Groupings
Since Google Analytics 4 doesn’t automatically capture and bucket AI search traffic, we’ve got to take matters into our own hands for now. Revisit your property setup to ensure LLM and AI traffic is properly bucketed.
Here’s how we did it:
- Open GA4 → Reports → Acquisition → Traffic Acquisition and add a Session source comparison to inspect the sources from which the traffic is actually coming from. If you see LLM domains like chat.openai.com or perplexity.ai, you’ve got LLM traffic coming in! This helps you build a list of sources to target.
- Go to Admin → Property column → Data Settings → Channel Groups. You can create a new custom channel group or copy the default (we recommend doing the latter so that you preserve existing logic and save yourself some extra work). Give it a clear name like “LLM & AI Traffic” and a short description.
- Click + Add condition group and choose the matching field Source (or Session source / source platform, depending on your UI) and set the operator to “matches regex”.
- Paste a tested regex of known AI referrers. You can expand it as you discover more sources, but start with the most common LLMs, like this:
- (chat\.openai\.com|chat\.openai|perplexity\.ai|perplexity\.com|gemini\.google\.com|claude\.ai|claude\.anthropic\.com|copilot\.microsoft\.com|chat\.google\.com)
- Order matters in GA4, so place your new LLM channel high in the channel rules list. Channel rules are evaluated top → bottom. Move the new channel above generic Referral or Unattributed channels so that your new rule captures sessions before they fall into those buckets.
- Save the channel and confirm that the GA4 UI shows your new channel.
Don’t Wait for the Playbook to Be Rewritten for You
Breathe in. Breathe out. Great; I know I just threw a lot at you, but thanks for hearing me out. All of this to say, SEO feels chaotic right now, and honestly, it kind of is.
For those of us who have dedicated our entire careers to playing nice with our dearly beloved Google, it may feel like the end. I like to think of it as an evolution. The metrics are changing because search itself is changing.
As a collective that prides itself on being “ahead of the game,” marketers who cling to outdated KPIs will misread the moment; those who adapt will lead the next generation of search strategy.
The TL;DR: Stop chasing what worked before and start measuring what matters now.
