Table of Contents
Google's AI overviews are fundamentally changing how SEO works, killing traditional top-funnel strategies while creating massive opportunities in mid-funnel search. Eli Schwartz, who's built SEO strategies for Zapier, Tinder, and Coinbase, reveals why most companies are approaching SEO completely wrong and what they should do instead.
The shift isn't just about AI answering questions—it's about understanding that most SaaS companies shouldn't even be doing SEO, and those that should need to completely rethink their approach to focus on where actual conversions happen.
Key Takeaways
- AI overviews are destroying top-funnel SEO opportunities, making long-form content about generic topics largely obsolete
- Mid-funnel SEO, where users have specific intent and problems to solve, represents the real opportunity for most businesses
- SEO should be treated as a product, not a marketing tactic—requiring product managers, not just marketers, to drive strategy
- Most B2B SaaS companies shouldn't do SEO because their customer journey doesn't align with search behavior
- Successful SEO requires solving specific user problems at the moment they're searching, not creating generic content
- Programmatic SEO (data-driven pages) typically outperforms editorial content when executed correctly
- Google maintains 98% market share on mobile and won't be displaced by ChatGPT or other AI tools anytime soon
- Keyword research tools are often wrong by factors of 10x, making top-down market sizing more accurate than bottom-up forecasting
- Technical SEO fixes won't save companies from Google algorithm updates—the problem is usually content quality, not technical issues
Timeline Overview
- 00:00–11:34 — AI's SEO Revolution: How Google's forced response to ChatGPT created AI overviews, the monetization challenges, and why this represents the biggest shift in search since Google's inception
- 11:34–22:57 — The Discovery Paradigm Shift: Why top-funnel searches now get AI answers instead of driving traffic, how user journeys are changing, and the rise of mid-funnel as the new SEO opportunity
- 22:57–39:17 — Product-Led SEO Framework: Zapier and Tinder case studies showing how to build SEO as a product, the importance of solving specific user problems, and why most SaaS companies approach SEO wrong
- 39:17–52:09 — The Economics of SEO Investment: How to calculate whether SEO is worth the million-dollar annual investment, why many companies should skip SEO entirely, and how to evaluate channel trade-offs
- 52:09–59:37 — Measuring Success and Timeline: Setting realistic expectations for SEO results, the difference between rankings and business metrics, and why conversion tracking matters more than traffic
- 59:37–1:09:51 — AI Tools and Brand Strategy: How to use AI for content creation without falling into the content trap, getting featured in AI overviews as brand building, and why programmatic beats editorial SEO
- 1:09:51–1:26:28 — Google's Monopoly Insights: Key findings from the 286-page antitrust verdict, why Google's 98% mobile market share won't change, and how distribution agreements create unbreakable competitive moats
- 1:26:28–1:46:26 — SEO Myths and Forecasting: Debunking link-building myths, why keyword research tools are often 10x wrong, and the top-down approach to forecasting SEO opportunities
The AI Overview Revolution: How Google's Hand Was Forced
Google's introduction of AI overviews wasn't a natural evolution—it was a panicked response to ChatGPT's threat to their search monopoly. "Google has claims to have invented the concept of LLMs, but ChatGPT came on the scene with this ability to ask any question and get a written out answer, and suddenly people are like, 'well I don't need Google anymore,'" Schwartz explains.
This forced Google into an uncomfortable position. They had to satisfy Wall Street's concerns while solving three major problems: monetization (ads don't work with AI answers), liability (Google becomes the publisher of potentially harmful advice), and plagiarism concerns (summarizing content without proper attribution).
The initial launch was embarrassing—screenshots of Google suggesting users put glue on pizza or jump off the Golden Gate Bridge went viral. Google rolled back the feature but has since relaunched it more aggressively. "I'm seeing it on most of the search results that I see, and it's now on non-logged in users and just launched in the UK," Schwartz notes.
This isn't just another Google update—it's a fundamental shift in how information discovery works online. The old model of ranking first for broad keywords is becoming obsolete as AI provides direct answers to top-funnel queries.
The Death of Top-Funnel SEO and Rise of Mid-Funnel Strategy
Traditional SEO focused on ranking for broad, high-volume keywords at the top of the marketing funnel. A travel site might target "best beach vacations within two hours of the United States" with long-form content designed to capture initial interest.
AI overviews have made this strategy largely obsolete. "You can ask a very explicit question about 'give me a beach vacation that is not in America but is a 2-hour flight from X airport'—that's the kind of thing that you could do in a Google search but would take you a very long time to do, and those are great answers from just getting a paragraph," Schwartz explains.
The new opportunity lies in mid-funnel search, where users have moved beyond general discovery to specific intent. After AI tells you to consider Cancun, you search for "Cancun hotels beachfront" or "Cancun flight deals from Boston"—these are the searches that still drive conversions.
This shift benefits users who no longer need to sift through SEO-optimized articles written by freelancers with no travel expertise. Instead, they get curated information that leads them to more targeted searches where businesses can actually convert them.
The Zapier and Tinder Case Studies: SEO as Product Strategy
Schwartz's work with Zapier and Tinder illustrates how successful modern SEO requires treating it as a product, not a marketing afterthought. Both cases involved identifying specific user problems that search could solve.
Zapier's Integration Discovery Problem: When Schwartz met Zapier's founder Wade Foster, the company had a product that connected different software tools, but no one knew it existed. "People knew that they needed things to work together, they just weren't looking for Zapier," Schwartz recalls.
The solution was programmatic SEO at scale—creating pages for every possible integration combination. When someone searched "Gmail Salesforce integration," they found a Zapier page explaining exactly how to connect these tools. This created a flywheel effect where users discovered additional integrations they hadn't considered.
Tinder's Loneliness Solution Framework: Tinder presented a different challenge—how do you do SEO for a mobile app? The insight came from user research: "Tinder is a loneliness solving solution. You're lonely, you've gone to a new city, you don't know anybody, you'd like to solve your loneliness problem."
This led to location-based SEO pages. Searching "online dating in Dubai" would lead to a Tinder page with local dating spots and positioning Tinder as the solution for meeting people in that specific city. The key insight: they weren't competing with generic dating advice but solving a specific geographic problem.
Both cases succeed because they identified mid-funnel problems that search could solve, then built product solutions specifically for those search contexts.
Why Most B2B SaaS Companies Shouldn't Do SEO
One of Schwartz's most controversial insights is that most B2B SaaS companies shouldn't invest in SEO at all. The reasoning comes down to user journey misalignment and conversion friction.
The Mixpanel Example: Schwartz worked with Mixpanel early in his consulting career, following traditional SEO tactics—identifying keywords, creating content, building links. "It didn't work, and we were sitting there and I asked them to show me the user journey... I realized that there were other problems to this conversion."
The fundamental issues: Mixpanel is an enterprise product that requires company-wide integration, has a complex sales process, and costs significant money. "You don't just do a quick Google search and be like 'Oh analytics, yeah I'm going to tell everyone we got to do it and it's going to be live tomorrow.'"
Google Cloud's SEO Impossibility: During a Google Cloud interview process, Schwartz discovered the absurdity of SEO for certain enterprise products. "For Google Cloud, they only have two competitors—Amazon and Microsoft—and no one is going to do a search and be like 'Oh Google's number one for this term I searched, let me just go buy it.' There's a decision-making process, there's a committee."
The pattern is clear: when your product requires sales conversations, committee decisions, and significant implementation effort, SEO traffic rarely converts directly to revenue. The channel serves awareness at best, but other channels typically provide better ROI.
The Economics of SEO: Million-Dollar Investment Reality
SEO isn't free, despite popular misconceptions. Schwartz breaks down the real costs that most companies overlook when evaluating SEO as a channel.
Hidden Cost Structure: A typical SEO investment includes agency fees ($10,000+ monthly), internal product management resources, engineering support, content creation, design resources, and tooling costs. "You need a CMS that costs money, you need an engineer to support that costs money, you need potentially design, you need content—it can really add up quickly."
The total often reaches $1 million annually when fully loaded. This creates a crucial question: "If I invested a million dollars a year in SEO, do I expect to make back a million dollars a year soon? And if they took that exact same million dollars and put it into brand ads or influencer campaigns or traditional paid marketing, would you make that million dollars back faster?"
The Garden Tool Company Reality Check: Schwartz encountered a SaaS company serving gardeners that wanted to spend $15,000 monthly on SEO. When he asked how they currently acquired customers, they revealed attending gardening trade shows at $10,000 per booth. "Instead of spending $15,000 on SEO, you could go to all these shows for the exact same budget and get users who are interested, they're in-market, they try your tool out at the booth."
This calculation should drive SEO investment decisions: not whether SEO could work, but whether it's the best use of limited resources compared to other growth channels.
Measuring What Matters: Beyond Traffic and Rankings
Traditional SEO metrics focus on vanity numbers that don't correlate with business success. Schwartz advocates for business-aligned measurement that tracks actual value creation.
The Traffic Trap: Most companies celebrate SEO traffic growth without connecting it to revenue. "No one doing paid marketing will ever say 'this is how many times I'm number one on this search' or 'this is how many clicks I get'—there's really a 'we spend this and we're this efficient, this is what we drive from that.' SEO really needs that same rigor."
Conversion-Focused Metrics: The right metrics depend on business model and user journey. For SaaS companies, track MQLs or qualified demos from SEO traffic. For e-commerce, track revenue attribution. For media companies, track engagement metrics that drive advertising revenue.
The HR Marketplace Lesson: Schwartz worked with a two-sided marketplace in HR space that generated significant traffic from job seekers but only monetized from employers. "I suggested that they delete all the content that was on the wrong side of that marketplace because they didn't convert at all."
This highlights a crucial principle: traffic that doesn't convert to your monetization model actually wastes resources and dilutes focus from traffic that does convert.
AI Content Strategy: Tools vs. Solutions
The rise of AI content creation has created both opportunities and traps for SEO practitioners. Schwartz's framework distinguishes between helpful AI use and content spam.
The Productivity Enhancement Model: AI works best when augmenting existing content strategies rather than replacing human judgment. "If the content you were creating was pretty useful and now you're using AI to create really useful content for cheaper and better, of course you can use it."
Examples include e-commerce product descriptions, feature explanations for existing pages, and standardized content that follows proven templates. The key is using AI to scale proven content formats, not to generate entirely new content strategies.
The Content Spam Warning: Many companies use AI to generate thousands of blog posts hoping to capture search traffic. "Using AI content to write more fluff content that's not necessary would just be a waste of time." This approach typically fails because it doesn't solve specific user problems.
Google's Quality Focus: Google's documentation emphasizes helpfulness over creation method. "AI itself is not the problem—it's the helpfulness, the usefulness of the content that would be a problem." This reinforces the importance of user value over content volume.
Google's Unbreakable Monopoly: Insights from the Antitrust Verdict
Schwartz's analysis of the 286-page Google antitrust verdict reveals why Google's search dominance won't be threatened by AI competitors anytime soon.
Market Share Reality: The verdict confirmed Google's overwhelming dominance: "Google has 98% of mobile searches." This wasn't Google's estimate—it came from court documentation that Google didn't dispute.
The Distribution Advantage: Google's default partnerships with Apple and others create insurmountable competitive barriers. "The plaintiff was the Department of Justice, but one of the complainants was this company Neeva, which was a search engine started by a past Google employee who couldn't be successful despite having a better search engine because they didn't have any default distribution agreements."
The Revenue Share Reality: Google pays Apple billions annually through revenue sharing from search ads. "Bing tried to give itself to Apple for free, and Apple said there was absolutely no price that they would take to use Bing search within Apple." This demonstrates how Google's revenue model creates competitive moats that pure technology advantages can't overcome.
Habit and Brand Power: Even when users have alternatives, they choose Google. "Mozilla had a partnership with Yahoo and everyone switched back to Google." This suggests that even if regulatory action breaks default agreements, user behavior heavily favors Google.
The Future of Search: Why Choice Remains Essential
Despite AI's growing sophistication, Schwartz argues that search will remain essential because AI can't perfectly predict individual preferences and needs.
The Choice Problem: Current AI assistants provide single answers without options. "One of the reasons I think home assistants have never taken off is because you don't have choice—you talk to your Google and it gives you only one answer, you talk to Amazon and it gives you only one answer."
Personalization Limitations: Even advanced AI struggles with individual context and preference. "There's never going to be an AI that understands you so perfectly that it's going to know that for you personally, you would like to click result number five—that's the best fit for what you're looking for right now."
Action-Oriented Search: When users need to take specific actions—booking hotels, purchasing products, finding local services—search remains the optimal discovery mechanism. "The aggregate number of people needing to take that action and pay you for it does not change even if search volume gets cut."
This suggests that while information-seeking searches may migrate to AI, intent-driven searches that lead to transactions will remain in traditional search results.
Common Questions
Q: Should my B2B SaaS company invest in SEO?
A: Only if users search for specific problems your product solves and can convert without sales conversations. Most enterprise SaaS shouldn't do SEO because the buyer journey requires human interaction.
Q: How long does SEO take to show results?
A: It varies dramatically based on what you're building and your company's authority. Well-known brands can see immediate results from new pages, while startups may wait months for Google to notice and rank content.
Q: Are keyword research tools accurate?
A: No. They're often wrong by factors of 10x compared to actual Google Search Console data. Use them for relative comparisons and search intent research, not absolute volume forecasting.
Q: How do I get featured in Google's AI overviews?
A: Treat it as brand building rather than traffic generation. Focus on becoming a recognized authority in your space rather than optimizing specifically for AI inclusion.
Q: Will ChatGPT and other AI tools replace Google?
A: Unlikely. Google's distribution advantages, 25 years of data, and user habits create competitive moats that technology alone can't overcome.
Conclusion
The AI revolution in search represents both an end and a beginning for SEO. Traditional top-funnel content strategies are becoming obsolete as AI overviews handle general information queries. However, this creates opportunities for companies that understand the new paradigm: mid-funnel SEO focused on specific user problems and purchase intent.
Success requires treating SEO as a product discipline rather than a marketing tactic, with product managers driving strategy based on user journey analysis. Most importantly, companies must honestly evaluate whether SEO aligns with their customer acquisition model before investing significant resources in a channel that may not drive meaningful business results.
Practical Implications
• Audit your current SEO strategy: If you're targeting broad, informational keywords, AI overviews are likely reducing your traffic potential
• Shift focus to mid-funnel: Target searches where users have specific problems or purchase intent rather than general curiosity
• Evaluate channel fit: Honestly assess whether your customers actually search for solutions or discover them through other channels
• Invest in programmatic over editorial: Data-driven page creation typically scales better than individual content pieces
• Measure business metrics: Track conversions and revenue attribution rather than rankings and traffic volumes
• Use AI tactically: Leverage AI for content enhancement and scaling proven formats, not for generating entire content strategies
• Consider top-down forecasting: Base SEO opportunity estimates on market size rather than keyword research tool data
• Focus on user problems: Build SEO assets that solve specific problems in the moment users are searching, not generic educational content
