How Google built the web era's dominant platform — and survived the shift to mobile

Executive overview

Google entered the public markets in 2004 as a pure-play search business, then spent the next decade doing something investors hated: investing billions into products with no obvious connection to search. The real logic was strategic — every major product Google built or bought served either to grow web usage (and thus search queries), to neutralise Microsoft's control of the distribution stack, or to insulate the business from the next platform shift.

The through-line across Gmail, Maps, YouTube, Chrome, and Android is the same: protect and extend the ad auction that prints Google's money.

Google's product strategy as a defence architecture

• Every major product had a dual bottom line: revenue potential and strategic value against a platform threat
• Wall Street punished Google in early 2006 (stock fell 27%) for investing profits into non-search products — Google ignored them
• The core insight: Google's business lived on top of Microsoft's stack (Windows + Internet Explorer); any Microsoft move to favour Bing could destroy search revenue overnight
• The response was to make the web itself the platform — and make consumers addicted to web apps before Microsoft could react

Gmail and the birth of web applications

• Gmail launched April 1, 2004; the gigabyte of free storage was a 250x leap over competitors and appeared deliberately farcical
• Ajax (asynchronous JavaScript and XML) made Gmail feel like a native app in a browser — developer Paul Buchheit discovered the XMLHttpRequest API and built the first widely adopted Ajax application
• The invite-only launch (1,000 seed invites, each user given a handful to share) controlled infrastructure costs and created scarcity that drove viral adoption; Gmail invites sold on eBay for $150
• Content-matched ads against inbox keywords were the first version of AdSense and helped prove display advertising outside of search results
• Strategic value: Gmail gave Google a persistent logged-in relationship with users — its first real foothold in identity

Google Maps and the Web 2.0 unlock

• Google acquired Where 2 Technologies (two brothers in Australia) in 2003; the founders rewrote their desktop mapping app as a web app in three weeks, independently rediscovering Ajax
• Launched February 2005 — the MVP covered only North America; Europe was an ocean
• The Maps API (released 2006) triggered the Web 2.0 mashup era and made companies like Zillow, Uber, Airbnb, and Foursquare possible
• Maps now has over 2 billion active users; revenue estimates exceed $5–10 billion annually

Google Docs and Sheets — attacking Microsoft where it hurts

• Real-time multi-user collaboration was the only feature that couldn't be replicated in installed desktop software — it was a uniquely web-native capability
• Google acquired Writely (Docs) in March 2006 and hired the founder of the spreadsheet startup that became Sheets
• Nobody else could subsidise this: Google didn't need Docs to make money; it only needed Microsoft to be forced to bring Office to the web at great expense
• Result: Google has the majority of global users in productivity software; Microsoft still has the vast majority of dollars — both companies are satisfied with this arrangement
• Google's incremental infrastructure cost for Docs was negligible; real-time collaboration ran on the same systems as Search

YouTube — the best acquisition in tech history

• Founded 2005 by three junior PayPal employees; original tagline was "Tune In, Hook Up" (a video dating service)
• Three structural advantages: easiest upload (no copyright checks, immediate live), best in-browser video player, and embeddable player that made every third-party site a distribution channel
• Google bought YouTube in November 2006 for $1.65 billion in stock — a decision the CFO immediately regretted because Google's stock has since 20x'd
• After acquisition, YouTube lost ~$1 billion per year and made $30 million in revenue; the CFO quietly explored selling it
• The turnaround came from: mobile (more logged-in users, better personalisation), shifting the internal metric from views to watch time, and the 50/50 creator revenue share that built an entire creator economy
• YouTube 2024: $36 billion in ad revenue, over $50 billion including subscriptions (YouTube Premium, NFL Sunday Ticket) — now the second-largest media company by revenue after Disney, ahead of Netflix
• Estimated operating income: ~$8 billion; Moffat Nathanson estimated standalone value at ~$500 billion
• The strategic bonus: Google whiffed on social, but social media pivoted toward video — which Google owns

DoubleClick — plugging into Madison Avenue's money flows

• Google acquired DoubleClick in April 2007 for $3.1 billion in cash, beating off bids from Microsoft, Yahoo, and AOL
• DoubleClick had invented the programmatic ad exchange — the infrastructure that routes advertiser spend across publishers in real time, deeply integrated into agency financial systems
• Without DoubleClick, Google was locked out of premium brand advertising and the dollars flowing through large agencies
• Microsoft, blocked from buying DoubleClick, immediately paid $6 billion for the number-two player (Aquantive) — twice the price, half the value
• Neil Mohan, DoubleClick's head of product, later became CEO of YouTube; many consider him the most valuable person in the acquisition

Chrome — liberating Google from Microsoft's browser

• Google secretly began preparing for a browser in 2001 but Eric Schmidt delayed, not wanting to "moon the giant" while the company was young
• Instead, Google funded Mozilla and Firefox as a proxy, then hired key Firefox engineers to form an internal "product clients" group — led from McKinsey by Sundar Pichai
• Chrome launched September 2008 with six core innovations: V8 JavaScript engine (fast Ajax), per-tab process isolation (crashes don't cascade), sandboxed security, minimal chrome UI, the Omnibox (one bar for URLs and search), and WebKit rendering
• Growth trajectory: 40 million users in 18 months → 200 million by 2012 → 70% market share today
• Chrome OS (Chromebooks) won education but not the broader PC market; Chromium open-sourced gave Google leverage even when others forked it
• Counterfactual: without Chrome, Microsoft could have made Bing the default search engine on 70% of browsers the moment it launched

Android — surviving the mobile platform shift

• Google bought Android Inc. for $50 million in July 2005; Andy Rubin's team had been building an open-source OS originally aimed at digital cameras
• Motivation: Google could see mobile search traffic growing and knew that if smartphones ran Microsoft or Apple software, Google's search revenue was at someone else's mercy
• The less-than-free business model: Android was free to OEMs and carriers, and Google also paid rev-share on searches that originated from Android devices — making it economically superior to paying Microsoft for Windows Mobile
• iPhone was revealed in January 2007; the Android team scrapped the keyboard-based "sooner" prototype overnight and pivoted entirely to touchscreen
• The Motorola Droid launch on Verizon in holiday 2009 was the true inflection: Verizon marketed it aggressively as the anti-iPhone, and Android market share went from ~5% to 30% in one year, then 50% by 2011 and 80% by 2013
• Google pays $20 billion/year to Apple for Safari default search placement; it pays ~$10 billion/year to all Android OEMs and carriers combined — structurally far cheaper per device
• Play Store 2019 operating income: ~$7 billion; but the primary value of Android was always protecting search revenue through the mobile era, not the Play Store itself
• Google is one of the only companies in tech history to dominate in two successive computing paradigms (web era → mobile era)

Google Plus — the Vista moment

• Launched June 2011 as a top-down, command-driven response to the strategic threat of Facebook becoming the starting point of the internet
• Built uncharacteristically without a core technical insight — essentially a Facebook clone, launched desktop-first, with "circles" as the key UX metaphor
• Forced integration across all Google products: YouTube comments became Google+ posts, mobile ad units got "+1 buttons," company-wide bonuses were tied to Plus metrics
• While Google built Plus, Facebook bought Instagram and WhatsApp and pivoted toward what would become the winning paradigm — public media and private messaging
• Plus shut down in 2019; the real survivors were Google Meet (from Hangouts) and Google Photos (over 1 billion users today)
• Biggest costs of the distraction: missing messaging (Google should have bought WhatsApp), underinvesting in cloud (Google Cloud remained years behind AWS and Azure), and a cultural slowdown that persists as a complaint about Google's product velocity today
• One genuine gain: forced company unification around a single Google account/identity — the fiefdom era ended

The Alphabet reorganisation

• In August 2015, Google restructured into Alphabet, with Larry Page as Alphabet CEO and Sundar Pichai as Google CEO
• Sundar had led Chrome and Android — the two platforms that protected Google through its biggest existential threats — and was seen as a unifier after the Plus era
• "Other Bets" broken out: Waymo, Nest, Google Fiber, Calico, Verily, Google X, Google Ventures
• 2015 financials: $75 billion in total revenue, $52 billion from Google first-party properties, $23 billion operating income; Other Bets lost $3.5 billion
• The core business in 2015 was — and still is — search advertising; the entire decade of product launches was infrastructure to keep that machine running

The AI groundwork

• Between 2015–2016, Google employed virtually every figure who would go on to define the AI era: Jeff Hinton, Ilya Sutskever, Dario Amodei, Andrei Karpathy, the co-founders of DeepMind, and the authors of the Transformer paper (published June 2017)
• The Transformer — the architecture underlying every major LLM today, including GPT — was a Google invention
• Larry Page articulated in 2000: "Artificial intelligence would be the ultimate version of Google… the ultimate search engine would understand everything on the Web"
• Google's decade of product launches also assembled the data assets (YouTube's video corpus, Maps, Search index, Gmail) that would become critical for AI training

Google's seven powers

• Counter-positioning: Android's less-than-free model was structurally impossible for Microsoft to match
• Scale economies: single advertiser buy across search, display, and video; lowest-cost infrastructure per query
• Network effects: YouTube two-sided creator/viewer flywheel; Android OEM/developer ecosystem; search auction liquidity
• Switching costs: 20 years of Gmail history; YouTube algorithm personalisation; Google account lock-in across products
• Branding: Google brand drove adoption of every new product launch; consumers sought Google Wave invites for a product that failed
• Cornered resource: YouTube video corpus for AI training; internal infrastructure (Borg) superior to anything available externally
• Process power: ability to launch and operate consumer-scale products at a fraction of what it would cost any competitor

Key playbook observations

• Google products that succeeded all had a core technical insight that was directly the user experience (PageRank, Ajax, real-time collaboration, video-at-scale, V8, mobile Linux) — products that lacked this (Google Plus, Wave) consistently failed
• Google's indirect business model — spend heavily on infrastructure, give away the products, earn from the resulting ad auction — is only possible with a money printer underneath; AdWords subsidised everything
• Google won two successive platform eras (web, mobile) by treating each platform shift as an existential threat early, not waiting for it to arrive
• The hit-factory era (2001–2015) was driven by a thin layer of tight strategy at the top and near-complete decentralisation below it — teams were told to build great products without needing to understand the ad business