Closing notes on 2025: what mattered in search
Core updates kept coming, AI features kept spreading, and the fundamentals kept winning anyway. An opinionated year-end review of SEO in 2025.
Year-end lists are a hazard of the trade, so here’s ours — short, opinionated, and centered on what we actually saw in ranking data rather than what got the most conference keynotes.
The update cadence is the new normal
2025 delivered core updates on a steady drumbeat — March took two weeks to roll out, June stretched nearly seventeen days into July — continuing 2024’s pattern (which packed four core updates into one year, two of them practically overlapping in November–December). The era of “wait for the algorithm to settle” is definitively over; the algorithm is the churn. Sites that treat measurement as a continuous practice rather than a post-incident forensic exercise had visibly easier years. We keep saying this; 2025 kept agreeing.
AI in the results: renegotiation, not apocalypse
AI-generated answers spread through more query types this year, and the discourse split between “SEO is dead” (again) and “nothing matters” (also again). The data we watch suggests something duller: informational head terms lost some clicks to answers-on-the-page, while specific, comparative, and transactional queries — the ones with money attached — still resolve in clicks. The lesson isn’t new, it’s the SERP-features lesson at scale: measure visibility and clicks separately, per query type, and skip both funerals and victory parades.
What quietly worked all year
The unfashionable list, unchanged: pages that answer one intent thoroughly; titles that say what the page is; sites without self-inflicted duplication; fast, crawlable, technically boring foundations. Every update cycle redistributes attention, and every cycle, the redistribution favors roughly this list. It’s almost as if Google means it.
What we’re watching in 2026
Whether AI features expand into commercial queries with the same appetite; whether update frequency keeps climbing; and whether “topical authority” keeps hardening from conference-word into observable ranking behavior (June’s update made that case strongly).
Our resolution is the same as last year’s: measure consistently, annotate honestly, panic never. See you in January.