Who invented the original Google search algorithm?

Larry Page and Sergey Brin co-authored the 1998 paper 'The Anatomy of a Large-Scale Hypertextual Web Search Engine' introducing PageRank. PageRank measured a page's importance recursively based on the importance of pages linking to it. The algorithm replaced earlier keyword-frequency approaches and became the foundation for Google's launch the same year.

What was BERT and what did it change?

BERT (Bidirectional Encoder Representations from Transformers) was deployed in 2019 as a bidirectional language model applied to query understanding. It improved Google's interpretation of complex queries — especially those with prepositions and conjunctions where meaning depends on surrounding words. Affected roughly 10 percent of queries at launch.

What is the Helpful Content Update?

Launched 2022. Targeted content created primarily for search engines rather than humans. Operates as a site-wide signal: a high proportion of unhelpful content can degrade rankings across the entire site. Forced a shift away from programmatic SEO content farms toward genuinely useful content.

When did Google add 'Experience' to EAT to make EEAT?

December 2022. Google added the first 'E' (Experience) to what had been EAT (Expertise, Authoritativeness, Trustworthiness) for years. Experience emphasized first-hand involvement with the topic — content written by someone who has actually done the thing tends to be more useful than content written by someone who only researched it.

What are AI Overviews and when did they launch?

AI Overviews are Google's AI-generated answer summaries that appear at the top of search results. Initially launched as Search Generative Experience (SGE) in 2023, rebranded to AI Overviews in 2024. Coverage expanded across query types through 2025-2026 and shifted some informational traffic away from traditional organic results.

The History of Search Engine Algorithms

Q: What was the Panda update?

Panda (launched February 2011) was a Google algorithm update that targeted thin content, content farms, and pages with high ad-to-content ratios. Panda updates ran every few months for years and forced the SEO industry to focus on content depth and quality rather than scale.

How search engine algorithms evolved from Google's 1998 launch through the AI search era. PageRank, Caffeine, Panda, Penguin, Hummingbird, RankBrain, BERT, MUM, Helpful Content, AI Overviews.

By David Schaefer · LinkedIn · Updated May 2026

Why this history matters

Search engine algorithms have evolved through roughly six eras since 1998. Each era introduced new ranking signals, penalized different patterns of manipulation, and reshaped the SEO discipline. Understanding the history clarifies why current ranking factors look the way they do — and why most "SEO secrets" you'll read about are remnants from an earlier era.

Era 1 — Link-based ranking (1998-2010)

1998: PageRank

Larry Page and Sergey Brin's 1998 paper introduced PageRank — a recursive measure of a page's importance based on the importance of pages linking to it. Inbound links functioned as votes. The algorithm was elegant, scalable, and significantly better than the keyword-frequency approaches of AltaVista and Lycos. Google launched commercially the same year. For roughly a decade, SEO was largely about acquiring backlinks.

2003-2009: spam fights

Florida (2003), Allegra (2004), Big Daddy (2005), Vince (2009) — a series of updates targeted keyword-stuffed pages, link farms, and low-quality content. Most were named by SEO commentators after their announcement timing.

Era 2 — Quality enforcement (2010-2013)

2010: Caffeine

Infrastructure update. Google rebuilt the indexing system to process new pages roughly 50 percent faster. Caffeine set the technical foundation for everything that followed.

2011: Panda

Targeted thin content, content farms (Demand Media, Associated Content), and pages with high ad-to-content ratios. Panda updates ran every few months for years. Forced the industry to focus on content depth and quality.

2012: Penguin

Targeted manipulative link-building — directory submissions, link networks, paid links, anchor-text manipulation. Made the previous decade's link-building tactics actively harmful.

Era 3 — Semantic understanding (2013-2018)

2013: Hummingbird

A major rewrite of the core algorithm to understand the meaning behind queries, not just the keywords. Marked the shift from keyword matching to entity and intent matching. Voice search and conversational queries became more practical.

2014: HTTPS as ranking signal

Google announced HTTPS as a lightweight ranking signal. The full migration of the web to HTTPS took years.

2015: Mobile-friendly update ("Mobilegeddon")

Sites without mobile-friendly design lost mobile ranking. Forced the responsive design transition for the long tail of the web.

2015: RankBrain

First major use of machine learning in core ranking. RankBrain helped Google interpret never-before-seen queries (roughly 15 percent of daily queries) by mapping them to similar previously-seen queries.

Era 4 — Neural understanding (2018-2022)

2018: Medic update

Targeted YMYL (Your Money or Your Life) sites. Health, finance, and legal sites saw major ranking shifts. Introduced the concept of expertise, authority, and trust (the original EAT) as evaluation criteria.

2019: BERT

Bidirectional Encoder Representations from Transformers. A bidirectional language model applied to query understanding. Improved Google's interpretation of complex queries, especially those with prepositions and conjunctions where the meaning depends on surrounding words. Affected roughly 10 percent of queries at launch.

2021: MUM

Multitask Unified Model. A multimodal, multilingual model that Google said was 1,000 times more powerful than BERT. Used for complex queries that require synthesis across multiple documents.

2021: Core Web Vitals update

Page experience signals — LCP, FID (later replaced by INP), CLS — became part of the page experience ranking factor. The technical SEO emphasis on speed and stability intensified.

2022: Helpful Content Update

Targeted content created primarily for search engines rather than humans. Site-wide signal: a high proportion of unhelpful content would degrade rankings across the entire site. Forced a shift away from programmatic SEO content farms toward genuinely useful content.

Era 5 — AI-generated content (2023-2024)

2023: Spam updates targeting AI content

Multiple spam updates targeted low-quality AI-generated content. Google's stated position remained "content quality matters, not how it was produced" — but in practice, much AI-generated content was thin and demonetized.

2023-2024: Search Generative Experience (SGE) / AI Overviews

Google launched generative AI responses at the top of search results. Initially called SGE, rebranded to AI Overviews in 2024. Shifted some informational query traffic away from the organic results below.

2024: EEAT (the second "E" added)

Google added "Experience" to EAT, making the framework EEAT. First-hand experience became a more explicit quality signal alongside expertise, authority, and trust.

Era 6 — AI search as default (2025-2026)

AI Overviews expanded across more query types. Perplexity, ChatGPT, Claude, and Gemini became significant sources of traffic and brand discovery. The discipline of optimizing for AI search engines (AEO and GEO) emerged as a distinct sub-field of SEO. Google's search results now compete with — and incorporate — AI-generated synthesis.

What persists across eras

Real content from real authorities ranks. Manipulation works briefly, then breaks when the algorithm updates.
Backlinks from authoritative sources remain a strong signal across every era.
User-experience signals (page speed, mobile, dwell time) became progressively more important.
Each era penalized the manipulation tactics of the previous era.
The trend has been from explicit signals (keyword density, anchor text) toward implicit signals (semantic meaning, behavior, trust).

What to read next

Sister pages: History of organic social algorithms, EEAT, Technical SEO, AEO and GEO.