We run SEO for 200+ clients. SMB dental, healthcare marketplaces, B2B SaaS, e-com, local services. Different verticals, different regions, different budgets. Same delivery pipeline. Every quarter we pull the data and look for the patterns. This is what 2024-2025 actually looked like.
The headline: SEO is not dead, but the playbook changed
Google's AI Overviews now appear in roughly 18% of commercial queries in our dataset. CTR on position #1 dropped from 27% to 19% on queries with AI Overviews. That's a real number, not a doom post. But the businesses that adapted — content structured for citation, schema, E-E-A-T signals that AI can read — saw position #1 traffic hold steady while everyone else bled.
Time-to-rank: faster than the agencies tell you
The agency-pitch timeline is 6-12 months to see results. Our actual median across 200+ clients: 78 days to first-page ranking, 142 days to top-3 for commercial queries. That's for sites with reasonable technical foundations and content production at scale.
The clients that took 6+ months had one of three problems: technical foundation was genuinely broken (noindex storms, redirect chains, JavaScript rendering issues), content production was bottlenecked on a single human writer, or the niche was so competitive that even a 10x effort showed up as 0.5x movement in the SERP.
AI content: the 80/20 nobody talks about
AI-drafted first versions, human-refined final versions, beats pure-human content on time-to-publish by 8x and on average ranking position by 1.4 spots. Pure AI content (no human refinement) loses to pure human content on every metric we tracked, including indexing velocity. Pure human content (no AI assistance) is now a luxury most agencies can't afford at scale.
The pattern that wins: AI for the 80% of the work that's repetitive (research, first drafts, schema, internal linking suggestions). Human for the 20% that matters (positioning, voice, claims, E-E-A-T signals, the parts an AI would get wrong). The math works because human time on the 20% is the part that actually moves rankings.
The metrics that predict revenue (and the ones that don't)
Traffic is a vanity metric. We tracked every metric against the client's actual revenue contribution from organic. The ones that correlated (r > 0.6):
- Commercial keyword positions in top 3 (not traffic, not impressions)
- Branded search volume growth (signals brand authority compounding)
- Pages indexed with rich results / FAQ / HowTo schema
- Referring domain growth rate (not absolute count, the velocity)
- Time on page for commercial pages (not blog posts)
The ones that didn't correlate (r < 0.3): total traffic, total impressions, average position, total backlinks, total keywords ranked. Those are great for client reports. They don't predict revenue.
What we're watching in 2026
- AI Overview expansion — currently 18% of commercial queries, projected 35-40% by end of 2026
- Zero-click search — already 65% of queries on mobile, climbing
- Brand authority signals — branded search growth is the leading indicator of compounding SEO wins
- First-party data — the brands that own their email list, podcast, and audience compounds SEO wins; the ones that don't, decay
- AI citations — the new 'rankings' are being cited in ChatGPT, Perplexity, Claude. The optimization is the same: structured data, E-E-A-T, content that AI models can read and quote.
The bottom line
SEO is more competitive than ever and more rewarding for the businesses that do it right. The playbook is: ship technical foundation fast, scale content with AI + human, measure commercial outcomes not vanity, and build brand authority as a compounding moat. That's what works. The agencies still selling 25 hours of human labor per client at $5K/mo are the ones being replaced.
If you want to see the underlying dataset — our 200-client benchmark spreadsheet with anonymized verticals, regions, and metrics — get in touch. We share it with agencies and direct clients on request.