On Google

• In 2019, Alphabet earned $162B of revenue
  • Of that $162B from Alphabet, $161B was Google (99%)
  • Of that $161B from Google, $135B was ads (84%)
  • Of that $135B from ads, $113B was ads on Google-owned websites (84%)
  • And of that $113B in Google ads, around ~$76B was phone ads (~67%)¹
• Of that $162B in revenue, 79% was plowed back into the business
  • $30B in traffic acquisition costs (19%)
  • $42B in other costs of revenues (26%)
  • $26B in R&D (16%)
  • $18B in sales & marketing (11%)
  • $10B in general & administrative (6%)
• After operating expenses, Alphabet earned $34B
• Per employee (~110K), this works out to be $1.5M in revenue, $1.2M in operating expenses, and $0.3M in operating earnings
• Per user (~1-2B), this is something like ~$100 in revenue, ~$80 in operating expenses, and ~$20 in operating earnings.
  • What’s more, if we assume that that ~$100 in ad revenue was for products with 20% margins, it implies Google’s ads cause each of us to shift our annual spending by an average of $500+. Wow.
• The reason Google can pay employees so much while getting only $100 per user is scale: for each employee, it has ~10K users
• Every Google ad you’ve ever clicked is shown here: https://myactivity.google.com/myactivity?hl=en&st=ad&product=27
  • I apparently click ~1 ad a month (26 clicks over last 24 months), despite feeling that I almost never click on ads. (I can’t recall how many of those were intentional vs unintentional, or if any led to purchases.)
• One fact I’ve never internalized enough is just how quickly these tech companies - even the giant ones - continue to grow. Like, you’d think that after 20 years, the search market would be pretty mature, right? Who hasn’t heard of Google? But no, Google revenue has been growing around 25% a year since 2009. Ad clicks grew 46% in 2017, 62% in 2018, and 23% in 2019. Search is not yet saturated!
• When an organism grows, each new cell inherits the DNA of the preceding cells. When an organization grows, you can only hope that each employee inherits the culture of the preceding employees. At the end of 2019, Google had ~120K employees, twice what it had four years before, which itself was twice what it had four years before that. Over the past 10 years, Google has grown headcount an average of ~20% per year!
  • Incidentally, this explains a lot of terrible reporting that equated a low median employee tenure with high turnover. In fact, low median employee tenure was a sign of fast growth.
• One distinguishing feature of Google is its strategic choice to invest billions into products that it gives away for free (e.g., Android and Chrome). Google is able to do this because so much money goes through Google search. If Google can make the internet slightly easier to use, so that people use the internet slightly more, that is worth billions to Google. If Google can give away free products that reduce a competitor’s ability to steer users away from Google search, that is worth billions to Google. Search is Google’s golden goose, and if any product appears unprofitable or irrational on the surface (e.g., Google Fiber), your first hypothesis should be that it somehow complements search.
• Google has eight products with 1B users:
  • Android + Google Play Store
  • Chrome
  • Drive
  • Gmail
  • Maps
  • Search
  • YouTube
• Although Google is considered visionary, most of Google’s products have been late to market (see below). In some cases, Google proved that doing things well beats doing things first, even in markets with strong returns to scale. But in other cases (e.g., cloud), being late to market was probably costly.
  • Late to search (1998, 3 years after Alta Vista, Yahoo, etc.)
  • Late to free web-based email (2004, 8 years after Hotmail)
  • Late to internet maps (2005, 9 years after MapQuest)
  • Late to internet browsers (2008, 13 years after Internet Explorer and Opera, 6 years after Firefox, 5 years after Safari)
  • Late to smartphones (2008, 1 year after iPhone, 8 years after Pocket PC 2000)
  • Late to social networking (2011, 7 years after Facebook, 14 years after Six Degrees) [ignoring Orkut, which was founded/copied in 2004]
  • Late to cloud infrastructure (2011, 6 years after AWS)
  • Late to cloud storage (2012, 5 years after Dropbox, OneDrive)
• Many of Google’s best-known hits began outside of Google:
  • Maps (maps from Where2 Technologies, acquired in 2004; real-time traffic from ZipDash, acquired in 2004; satellite imagery from Keyhole, acquired in 2005)
  • Android (founded in 2003, bought for $50M+ in 2005)
  • YouTube (founded in 2005, bought for $1.7B in 2006)
  • Adsense (Google acquired Applied Semantics in 2003, DoubleClick in 2007, and AdMob for mobile ads in 2009)
  • Google Docs (spreadsheets by 2Web, acquired in 2005; docs by Upstartle, acquired in 2006; slides by Tonic Systems, acquired in 2007)
  • Waymo (Thrun, 510 Systems?)
• Google acquires companies frequently (more than 1 company per week in 2010 and 2011): https://en.wikipedia.org/wiki/List_of_mergers_and_acquisitions_by_Alphabet
  • One theory says that today’s tech giants have learned from the declines of their predecessors. IBM, Sun, Microsoft, etc. failed to fully anticipate seemingly small trends like the PC, the internet, and mobile phones, all of which ended up eclipsing their present market. Today’s tech giants have taken the lesson to heart and will acquire startups at a premium, where that premium is effectively financed by acting as an insurance policy on the tech giant’s core business.
  • Although this theory sounds reasonable, it’s difficult to falsify without sitting in the decision room with Google’s executives. Alternate theories include Google believing that (1) tech startups are systematically undervalued, or (2) that there are synergies from sharing talent, tech, distribution, and brand across a portfolio of products.
• Alphabet has ~14 subsidiaries
  • Google (~$10B profit per quarter)
  • Other bets (~$1B losses per quarter)
    • Access (fiber)
    • CapitalG (growth equity fund)
    • Calico (aging & age-related diseases)
    • Chronicle (computer security)
    • DeepMind (machine learning)
    • GV (VC fund)
    • Jigsaw (incubator/think tank?)
    • Loon (internet balloons)
    • Sidewalk Labs (urban tech - pilot plan for 12-acre waterfront in Toronto)
    • Verily (life science tech - glucose monitoring, smart contact lens)
    • Waymo (autonomous driving - pilot in Phoenix)
    • Wing (drone delivery - pilot in Australia)
    • X (experimental ideas)
• Already I’ve lost much memory of what the internet boom felt like. Back in the late 90s and early 00s, sites were slower and clunkier and less user-friendly. I remember how Google and other companies had a magic feeling of putting the user first. Things like smooth sign up flows, fast performance, non-corporate communication were big deals. Gmail, Maps, and Chrome were all cool when they came out. Blogs were cool. RSS was cool. Digg was a big source of internet culture content. Small individual websites seemed like more a thing. My experience of these internet years were certainly colored by being in high school and college at the time (2002-2010). Generally I think websites spread faster among teenagers, since teenagers have more free time to experiment, more social time to share, more curiosity, and less feeling of saturation.

Notes