The way we measure success in SEO has changed. With the introduction of AI Overviews and AI Mode, Google now answers a growing share of queries before a user ever reaches a website – and some won’t even get to Google if their question is answered by an LLM, like ChatGPT, first. As a result, many of the metrics SEOs have relied on for years need reevaluating.
The new SEO checklist: metrics that still matter
The era of chasing traffic for the sake of a graph is ending. The reality is that a brand can now exert influence and drive decisions in search without a click ever happening. This shift requires a total rethink of what we value. Instead of obsessing over volume, we need to focus on intent, authority, and the actual outcomes that keep a business moving forward.
Start where SEO was always meant to end: outcomes
If your reporting doesn’t ladder up to what the page is meant to achieve, it’s just noise.
Conversions, leads, downloads, sign-ups – whatever “success” looks like for that page – should sit at the top of the checklist. Everything else is a supporting signal, not the goal itself.
More traffic does not automatically equal more success. Ten thousand irrelevant visits are less valuable than a hundred people who actually need what you offer and take action. SEO has always been about intent, but AI-led discovery makes that distinction impossible to ignore.
If a page isn’t contributing to outcomes, it doesn’t matter how visible it is.
Traffic is optional. Relevance is not.
One of the most uncomfortable truths for marketers is that SEO can be “working” while traffic is flat or even declining.
In an AI-first SERP, fewer users are clicking through. That doesn’t mean your content isn’t influencing decisions. It means discovery is happening higher up the funnel, often without a visit.
This is where many teams panic and start chasing volume again. That’s the wrong reaction.
The better question is: are we attracting the right people when they do land? And are we giving them a clear, confident experience once they’re there?
Impressions matter more than before
For years, impressions were seen by some as a vanity metric. In an AI-led search landscape, they’ve become more strategic.
When users see AI answers, citations and source lists instead of traditional blue links, brand presence is the goal, rather than clicks. Being cited, mentioned or visually present in an AI Overview builds familiarity and memory, even if the user doesn’t click there and then.
This isn’t new marketing theory, it’s basic brand building.
Repeated exposure matters as a user might not need you right now because the AI answer solved their immediate problem. But, when your brand appears consistently as a source, it might be mentally short-listed for later.
You can’t optimise for clicks alone when discovery increasingly happens without them.
Engagement is a quality signal you can’t ignore
If traffic is less reliable, what happens after the click matters more. Time on page, scroll depth, interaction patterns and content consumption all give you clues about whether your page is genuinely useful. Tools like Hotjar, session recordings and heatmaps come in handy here.
Engagement tells you whether your content deserves to be cited, remembered and trusted.
If users bounce immediately, skim without engaging, or never reach key sections, AI engines won’t see your content as authoritative for long. Poor experience quietly erodes credibility.
SEO and UX are no longer separable disciplines – treat them as such at your peril.
Be sceptical (but not dismissive) of AI tracking tools
There’s an explosion of tools promising to tell you how often your brand is cited in AI answers: Profound, Peec AI, and Otterly just to name a few.
Used carefully, they can provide directional insight. Used blindly, they can create false confidence.
The reality is that AI answers are personalised and generative. Research shows that if 100 people submit the same prompt, they’ll receive 100 different answers – with less than a 1% chance of seeing the same list of brands in the same order.
No tool can definitively tell you how often you’re “ranking” in AI. Nobody has that data.
Treat these platforms as trend indicators, not sources of truth. They’re inputs for thinking, not reporting KPIs.
Rankings still matter – just not for the reason you think
Despite the noise, traditional search rankings are not dead.
Large language models rely on grounding – often through live Google searches – when generating answers outside their training data. Ranking well for relevant queries still increases your chances of being surfaced, cited or used as a source in AI-driven results.
So yes, keep tracking rankings. Keep measuring impressions. Keep doing the fundamentals of “standard SEO”.
Just stop pretending those metrics are the end goal. They are distribution signals, not success measures.
The bottom line
AI has changed how discovery works, but it has not changed why SEO exists. The fundamentals remain. The brands that win are not chasing every new metric or dashboard. They are focusing on visibility that builds memory, experiences that earn trust, and measurement that ties back to real outcomes. They’re optimising for authority, intent and results.
