Three years of Google algorithm updates: what actually moved rankings
By Kirk Musick, MS, MBA
May 2026 operator update
Current read: algorithm history matters only if it changes what you do this month. The March 2026 core update, Google’s AI Search source-link changes, and rising AI Overview visibility all point in the same direction: thinner pages lose, clearer expert sources become more valuable.
- What changed: SERP volatility now includes ranking shifts and answer-surface shifts.
- What to review: pages that lost impressions, pages that lost clicks despite stable rank, and pages that are not cited where they should be.
- Current sources: March 2026 core update completion; Google AI Search link update, May 2026; Ahrefs AI Overview visibility research.
Google announces 4–6 broad core updates a year, plus spam updates, helpful content updates, and product review updates. The SEO industry reacts to each one with high drama. Most of them don’t matter much. A few quietly reshape what wins.
Looking back across the last three years of update activity, three patterns separate the updates that actually moved rankings from the ones that produced more discourse than impact.
Pattern 1: Helpful Content was the real shift
The 2022 Helpful Content Update — and its continuous absorption into the core algorithm in 2023–2024 — did more to reshape what ranks than any other change in the period. Three durable effects:
Sites with thin AI-generated content lost ground. Not because Google has a perfect AI detector, but because the kind of content AI produces without an editorial layer reads as low-effort, low-original-perspective, and patterned. The Helpful Content classifier picks up the patterns. Sites that rolled out hundreds of AI-only posts in 2023 saw cliff drops; sites that used AI for outlines + human writers for finals held steady.
Author E-E-A-T became measurably important. Pages with named, credentialed authors gained relative to pages with no byline or generic “Editorial Team” attribution. This was always supposedly true; in the Helpful Content era, it became operationally true. Bylines, author bios, links to author social profiles, mentions of credentials — the structured signals of “a real person with knowledge wrote this” — moved rankings in YMYL (Your Money or Your Life) verticals especially.
Comprehensive “ultimate guide” content lost to focused content. The 3,000-word “everything about X” posts that ranked in 2020–2022 lost ground to 1,000-word posts that answered the specific query the searcher had. Word count stopped being a proxy for depth. Depth-of-answer became depth-of-answer.
The aggregate effect: the SEO content factories that dominated 2018–2022 lost ground to operators who’d been doing what Google had said all along — write for the searcher, not for the algorithm.
Pattern 2: AI Overviews changed which queries matter
This isn’t strictly an algorithm update, but it’s downstream of the same evolution. The rollout of AI Overviews and the expansion into AI Mode reshaped the value of different query types:
| Query type | Impact of AI Overviews | What to do |
|---|---|---|
| Definitional (“what is X”) | Answer panel often appears; CTR drops 18–64% | Don’t build content strategy here anymore |
| How-to / informational | Mixed — Overview sometimes appears, sometimes doesn’t | Optimize for citation within Overview rather than competing |
| Commercial investigation (“best X for Y”) | Overviews appear less often; when they do, they typically link to sources | Still high-value; optimize aggressively |
| Transactional (“buy X near me”) | Overviews rarely appear; local pack remains dominant | Highest-value query type for most businesses |
| Branded (“X company reviews”) | Mixed — sometimes Overview summarizes; sometimes blue links dominate | Watch carefully; manage brand reputation aggressively |
For most service businesses and local operators, this was net positive — informational queries got harder but those weren’t the high-converting queries anyway. The queries that drive revenue mostly didn’t get easier or harder; they just kept rewarding the operators doing the technical and content work properly.
Pattern 3: Spam updates kept doing what spam updates do
Across the period, Google ran multiple spam updates targeting:
– Link spam (low-quality backlinks, link networks, expired-domain manipulation)
– Scaled content abuse (programmatic AI content at volume)
– Site reputation abuse (the “parasite SEO” pattern where weak sites publish on strong domains)
– Cloaking and redirect manipulation (showing different content to Googlebot vs. users)
Sites doing any of this got hit. Sites not doing it didn’t. The spam updates aren’t reshaping legitimate SEO; they’re cleaning up cheaters. The takeaway for legitimate operators is just confirmation that the work that wins isn’t the tactical-manipulation work — it’s the foundational work.
What didn’t matter as much as the industry said it did
A short list of things that produced more SEO discourse than ranking movement:
- The MUM rollout (2021). Big announcement, mixed measurable impact.
- Page Experience as a ranking factor (2021). The Core Web Vitals component matters at the margin but doesn’t override content quality.
- Most “named” updates between 2018 and 2021. The Florida update, the Page Experience update, the Product Reviews updates — each shifted rankings for a subset of sites, but the aggregate effect was small.
If you tracked rankings against every announced update over the past three years, the variance from algorithm updates is smaller than the variance from your own publishing cadence, backlink acquisition rate, and competitor activity. The updates matter; they’re not the main variable.
The post-update playbook (the actual one)
When Google announces an update, the discipline most operators don’t follow:
- Wait 14 days minimum before drawing conclusions. Rollouts take that long to propagate and stabilize. Day-3 takes are noise.
- Segment your impact analysis. Did all pages drop or just some? Commercial vs. informational? One topic cluster or sitewide? The pattern tells you what to fix.
- Compare to competitors who didn’t drop. They’re the diagnostic — what do they have that you don’t?
- Make one change at a time. Bulk-changing in response to an update means you can’t attribute what worked.
- Trust the boring inputs. The post-update fix is almost never a new tactic. It’s usually finishing the technical hygiene, deepening the content, or improving the author signals that should have been there already.
What 2026 looks like
The pattern from the last three years suggests the same pattern continues: a few updates quietly raise the bar on what “good content” looks like, the SEO industry overreacts to most announcements, and the operators doing the foundational work compound.
The single biggest 2026 signal worth watching: AI inclusion (being cited in AI Overviews and AI Mode) is becoming a ranking-adjacent signal in its own right. Sites optimized for AI citation — strong schema, clear topical authority, named experts — are likely to gain rankings on related queries even when they’re not cited directly. That’s a continuation of the E-E-A-T trend, not a new thing.
Most updates don’t matter. A few quietly reshape what wins. The work is the same either way.
Operator summary
- Most algorithm volatility is easier to understand when grouped by page type and intent.
- Watch helpful content quality, technical trust, authority signals, and AI Overview impact before chasing tactics.
- AI/search signal: consistent expertise and structured answers make a site more likely to survive SERP format changes.
Related ZINC guides
ZINC Digital builds organic search programs for service businesses, mid-market e-commerce, and local operators in Miami and Panama City. We start every engagement with an audit, then move into a monthly retainer with weekly working sessions and monthly performance reviews — tied to revenue, not sessions.
