AI is reshaping marketing. 

If you’re a marketer, this is likely what you believe. Even if you don’t believe it, you’ve certainly heard the sentiment again, and again, and again over the course of the past few years.

While it’s certainly not an incorrect belief, the truth is less dramatic. And far more useful.

Rand Fishkin is - there’s no other word for it - one of marketing's GOATs. 

(For anyone in doubt, that’s ‘greatest of all time’. He doesn’t eat grass and hang out on precarious cliff edges - to our knowledge. 🐐)

We were fortunate enough to speak to Rand (CEO & Co-Founder of SparkToro, Moz, Alertmouse, and several other companies) on the Paid Media Lab podcast, to help us separate AI hype from reality and provide marketers with a sobering view on what marketing’s role will look like going forward.

Rand has spent decades analysing how audiences find and engage with information online. His current research at SparkToro tracks how behaviour shifts across search, social, and AI-driven tools like ChatGPT and Perplexity.

In this episode, Rand discusses his findings, which tell a clear story: 

Most of what’s being said about “AI taking over marketing” doesn’t hold up when you look at the numbers.

This episode breaks down:

• Whether we’re living through an AI bubble
• What’s really happening to search behavior and traffic
• Why organic visibility is declining across industries
• How marketers should adapt their attribution and measurement strategies
• What practical steps to take to stay ahead of AI-driven changes
  
  

Rand’s data-led perspective brings much-needed realism to the conversation. Instead of chasing hype cycles, he focuses on what still works, and what’s quietly changing underneath the noise.

Watch the full episode here, or keep reading for the written recap.

Episode timestamps:

0:00 - Coming up
01:25 - Intro 
04:45 - Is Rand an AI optimist or pessimist?
06:26 - Are we in an AI bubble (in marketing?)
08:48 - Is it the same as the Dot-Com crash?
10:13 - ChatGPT 5 & AI overhype
12:22 - Is Google dead? (+ Research)
17:48 - The attention economy
21:07 - Why web traffic has declined
26:04 - The marketing mindset shift
31:29 - Is attribution broken?
36:22 - Should you advertise on AI platforms?
39:47 - ChatGPT’s Instant Checkout
42:37 - AI platforms will screw you
43:45 - How to cheat LLMs
48:15 - What marketing leaders need to focus on now
52:19 - Final thoughts

The AI bubble and the reality behind the hype

Are we in an AI bubble?

If you’ve had eyes and/or ears at any point over the last few years, you’ll already know the answer is ‘yes’.

The question then becomes ‘how big is the bubble?’ and ‘what on earth is going to happen when it pops?’

Rand explained that the AI boom isn’t just shaping marketing. It’s holding up entire economies. 

In the US alone, more than a trillion dollars have already been spent on AI hardware. Public markets, venture capital, and corporate roadmaps are all tied to AI optimism.

That optimism creates pressure for marketers to act fast, even when the outcomes aren’t clear. Rand’s view is that this rush mirrors the dot-com era, where overinvestment and inflated promises led to a hard correction.

Rand noted that while AI tools have utility (summarising data, brainstorming copy, automating repetitive tasks) - they’re far from the transformative shift some predict.

In his words, AI today is, quote, a “spicy autocomplete.” 

It can accelerate certain workflows but can’t replace the fundamentals of good marketing: understanding audiences, creating compelling offers, and measuring impact.

Key takeaway: We are in an ‘AI bubble’. Use AI to enhance output, not to define strategy.

ChatGPT vs Google: Separating hype from data

Rand’s team at SparkToro recently analysed real user behavior across major search and AI tools:

Key findings:

• Google dominates search. It still drives roughly 210 times more searches per day than ChatGPT.
  
  
• ChatGPT’s usage is smaller than DuckDuckGo’s. Despite the buzz, it’s far from a mainstream search habit.
  
  
• Desktop vs mobile skews the story. Google’s strength lies in mobile, while ChatGPT remains primarily desktop-based.

Rand expands on the findings in more detail in the podcast, but he also created a standalone video discussing them here, which is definitely worth a watch. 

In short: AI chat tools are growing, but their influence is limited compared to traditional search.

Rand explained why this gap matters for marketers. When people claim that “ChatGPT is replacing Google,” they’re ignoring basic usage data. The truth is that Google continues to be the backbone of information discovery, while ChatGPT and similar tools act as secondary layers - useful for summaries or quick answers, but not the starting point for most online journeys.

He also highlighted how the “Google is dead” narrative spreads online. It benefits almost everyone except marketers trying to make smart decisions:

• Media outlets gain clicks from bold predictions.
• Agencies and consultants use fear to sell new AI solutions.
• Investors fuel inflated valuations for AI-first startups.

For marketers, this means staying grounded in user behaviour data instead of reacting to hype cycles. Search still underpins most intent-driven discovery. AI tools sit on top of that foundation, extending how people consume and explore - not replacing it.

The attention economy (and why hype always wins)

Rand broke down how modern media incentives amplify extremes. Moderation rarely travels far online, but fear and novelty do.

When it comes to AI, this means the loudest claims - good or bad - get the most visibility.

Key insights from the conversation:

• Attention bias drives perception. Calm, data-driven takes never go viral. Alarmist predictions do.
• Engagement algorithms reward outrage. Platforms prioritise emotional reactions over accuracy.
• The result: extreme narratives dominate discourse, while balanced perspectives get buried.

Rand explained that this same mechanism applies to marketing itself. Ad algorithms on Meta, Google, and TikTok are built on similar engagement loops. The content that sparks strong emotion or surprise tends to outperform carefully polished creative.

Rand’s advice is to understand the environment you’re operating in. When you know that algorithms thrive on extremes, you can shape messages that cut through without compromising credibility.

Key takeaway: Build strategies grounded in attention mechanics, not assumptions about truth or fairness. Marketing still runs on emotion and momentum, and AI hasn’t changed that.

Why your website traffic is falling, and what’s actually behind it

Many marketers assume AI is stealing their clicks. Rand’s data paints a different picture.

Traffic declines aren’t being caused by ChatGPT or Perplexity. They’re being caused by Google itself.

Google now answers over 60% of searches without a click. Between AI overviews, featured snippets, and its own ecosystem products (YouTube, Maps, Flights, Hotels), Google is keeping more users inside its walls.

Rand used HubSpot as an example.

Their organic traffic dropped by 80%, yet revenue and share price hit record highs. The reason? Their audience is finding and engaging with them elsewhere - on YouTube, LinkedIn, and other discovery platforms.

He calls this shift “zero-click marketing.”

It’s the idea that visibility now matters more than visits. Your job as a marketer is to influence your audience wherever they already are, not just drag them back to your website.

Learnings from Rand’s analysis:

Rand made the point that websites still matter - but their role has changed. They now act as sources of truth and credibility that feed other channels and AI systems.

“Your website influences the web more than it captures traffic from it.”

Key takeaway: Success comes from being seen, cited, and shared across platforms, even when that doesn’t generate an immediate visit.

The rise of platform-first marketing

Rand described a fundamental change in how marketing ecosystems operate. Ten years ago, everything revolved around bringing users to your website. Today, platforms dominate.

“Marketers used to optimize for traffic. Now they need to optimize for influence.”

Instead of fighting algorithms for clicks, Rand recommends meeting audiences where they already spend time. That means designing content and creative for native impact - within the feeds, forums, and search results that matter most.

Attribution is broken.  What to do instead:

Rand didn’t hold back here.

In his view, attribution as we know it is finished. Privacy changes, walled gardens, and inflated platform reporting have made the numbers unreliable.

He described how marketers have been conditioned to trust dashboards that give the illusion of precision. In reality, those dashboards are just showing where conversions are being credited - not where they actually started.

His favourite example was a hotel analogy. Someone might hear about a hotel through a podcast, Reddit thread, or friend recommendation. Later, they search the name on Google and make a booking. Google takes full credit, even though it had nothing to do with the discovery.

This same bias affects Meta and other platforms that reward themselves for conversions they didn’t cause. Rand mentioned companies like Airbnb and Chase that cut ad spend for months and saw almost no drop in bookings. The takeaway: most paid media platforms overreport performance.

Rand encouraged marketers to shift focus from attribution to measurement. Instead of trying to assign every sale to a single click, look at lift, patterns, and exposure over time. Reduce ad spend for a short period and see what really changes. Monitor geographic or time-based variations.

Measurement is about correlation and trend, not false precision. It’s slower and harder to communicate, but it gives a truer picture of where your marketing actually creates impact.

“Attribution isn’t dead. It just stopped telling the truth.”

What marketers should actually do about it

Rand’s message throughout the episode was pragmatic. The marketing fundamentals haven’t changed. What’s shifted is how and where those fundamentals show up.

He outlined five practical actions for marketers who want to stay effective in an AI-saturated landscape:

1. Relearn measurement

Forget the illusion of perfect data. Treat metrics as directional signals, not absolute truth. When you see a performance spike, question it. Test it. Understand what really caused it.

2. Double down on audience research

Before investing in any platform or ad channel, find out where your audience actually spends time. Use tools like SparkToro or Similarweb to map their habits and interests. Then allocate budget to those attention zones.

3. Adopt a zero-click mindset

People might see your message on YouTube, Reddit, LinkedIn, or in a podcast transcript before they ever hit your site. Influence them early. Be visible in the spaces that shape perception.

4. Test the real impact of paid media

Reduce spend temporarily. If conversions barely change, the platform was probably taking credit for demand created elsewhere. Keep testing small and often.

5. Bake marketing into your product

Rand’s strongest recommendation was to build products that market themselves. He gave the example of SparkToro’s new interface, redesigned to encourage sharing. Screenshots of data visualisations now circulate naturally on social media, generating organic reach without extra spend.

He closed the section by reminding listeners that great marketing has always been about visibility and talkability. AI tools might amplify the process, but they can’t replace creativity, curiosity, or cultural relevance.

“Make something people want to share. If your work spreads on its own, attribution becomes irrelevant.”

Getting your brand mentioned by AI

Rand peeled back the curtain on how large language models surface brands in their answers.

Right now, the process is crude. These systems rely heavily on recency and repetition. If your brand is mentioned often across the web - especially in public, crawlable formats - it’s more likely to show up in ChatGPT, Gemini, or Perplexity results.

He pointed out that many agencies are exploiting this gap by publishing dozens of keyword-stuffed microsites. They include phrases like “Best SEO agency in Toronto” to game LLM training data. It works in the short term, but it won’t last.

The smarter, sustainable route is to build real credibility. Earn legitimate mentions through podcasts, press coverage, YouTube videos, and high-quality blog content. These sources carry authority and will remain in training data even as algorithms evolve.

Key tip for visibility: feed the ecosystem. 

Keep your brand’s name, expertise, and unique language active across channels that language models pull from. Transcripts, Q&A appearances, and thought leadership pieces all increase the chances of being cited.

“If you want AI tools to recommend you, be the brand they learn from.”

The key principle is consistency. Influence is cumulative, and every mention adds another signal. Over time, those signals define how both search engines and AI systems interpret who you are and what you do.

Crucial advice for marketing leaders

Rand wrapped up the discussion with a wider perspective on leadership in the current marketing environment. Budgets are tighter, noise levels are higher, and internal pressure to prove ROI has never been stronger.

He advised marketing leaders to stop chasing algorithm updates and short-lived attribution models. Instead, focus on clarity, creativity, and credibility.

• Clarity: Know what your product stands for and who it serves. If your offer isn’t clear, no AI tool or ad platform can fix it.
• Creativity: Encourage your teams to experiment. Campaigns that surprise and entertain still win attention.
• Credibility: Build long-term trust through consistent value, transparency, and expertise.

Rand also emphasised collaboration. The most effective marketing teams break down silos between content, paid, product, and analytics. Each function strengthens the others when strategy is shared.

“Marketers who understand how influence works will always outperform those chasing perfect attribution.”

Final thoughts

This episode of the Paid Media Lab stripped away the noise around AI and performance marketing. We wanted to thank Rand for joining us, and for sharing evidence behind where the future of marketing is actually headed, not 

Core takeaways:

• The AI boom is very real, but also very overhyped.
• Google’s dominance remains unmatched - even now.
• Website traffic is no longer the main success metric.
• Attribution is broken, but measurement still works.
• Visibility now happens across platforms and inside AI systems.

Rand’s grounded, data-backed approach offers something rare in today’s marketing world: perspective.

Watch or listen to the full episode here.

Explore more:

• Follow Rand on LinkedIn
  
  
• Get Rand’s book, Lost and Founder
  
  
• Subscribe to the Paid Media Lab podcast
  
  
• Get your 14-day Lunio traffic audit to see exactly where your ad spend is going
  
  

The future of marketing isn’t about chasing hype, but rather about understanding behaviour, earning attention, and staying visible - wherever your audience decides to look next.