NGraph

Google: company structure analysis (Arche)

G

Google

Startup·USAUSA, 1998
58
/100

Society

8/10

Customer

8/10

Self

4/10

11 above the survival line

Survival 47Excellence 75

'The link graph itself is the ranking signal.' In a market dominated by Yahoo!'s directory and AltaVista's keyword search, Page and Brin commercialized PageRank — a mathematical counter-thesis born from a PhD program.

12 criteria

36/60

+

3 axes

20/30

+

Bonus

+2

=

NCG

58/100

Bonus: [Insight+Strategy] $1M sale to Yahoo! fell through, leading to the infrastructure path; Bechtolsheim's pre-incorporation $100K + 600K daily PV

Context

Larry Page and Sergey Brin had been running 'BackRub' on Stanford's network since 1996, registered google.com in September 1997. In August 1998, Andy Bechtolsheim (Sun co-founder) wrote a $100K personal check before the company was even incorporated — that catalyst led to Google Inc.'s formal California incorporation on September 4. The team rented Susan Wojcicki's Menlo Park garage (Santa Margarita Ave) at $1,700/month as headquarters, starting with 4 employees (Page, Brin, Craig Silverstein, Heather Cairns). By year-end the index served roughly 600K page views per day, but no revenue model existed — AdWords wouldn't launch for another two years (October 2000).

TelesTeles' Insight

Google in September 1998 scored NCG 36 — a four-person team in Wojcicki's garage, borrowing Stanford's compute. Bechtolsheim's $100K personal check, written before the company was even incorporated, was the catalyst that triggered formal incorporation. Famously, in 1998 Page and Brin offered to sell to Yahoo! for $1M and were turned down. At a moment when conventional wisdom said 'sell to Yahoo! and exit,' only Page and Brin articulated 'mathematize the link graph and search becomes humanity's infrastructure' — Insight=5. AdWords was two years out, revenue was zero. But $1,700/month garage burn + 600K daily page views of viral growth secured survival through the NCG rule's triple-max: Insight 5 + Timing 5 (dot-com + web explosion) + Team 5 (Stanford CS PhDs). The $2T+ that came twenty years later started here.

Key to survival

Triple max — Insight 5 + Timing 5 + Team 5. Bechtolsheim's pre-incorporation $100K check plus Stanford's exclusive PageRank license enabled zero-revenue minimal-burn operation. Yahoo!'s rejection of a $1M sale ended up opening the path to 'become infrastructure, not exit.'

12 Criteria (36/60)

StrengthsB Problem, C Solution, D Team, L Insight

WeaknessesF Moat, G Revenue, H Financial, J Exit

AMarket4/5

The 1998 web-search market was still forming. AltaVista, Yahoo!, Lycos, Excite, and Inktomi competed; monetization leaned on ads with low CPMs. The full TAM wasn't yet clear.

BProblem5/5

Web page count had reached ~200M and was exploding. Incumbents matched keywords or curated directories, and quality was degrading — users routinely went three pages deep without finding what they wanted.

CSolution5/5

PageRank is a mathematical algorithm that scores the link graph via a random walk, recursively computing 'quality' from inbound links. The experience improvement over AltaVista was a clear 10x in perceived precision.

DTeam5/5

Page + Brin (a genius Stanford CS PhD pair) + Silverstein (first full-time hire). Backed by advisors Hennessy (later Stanford president) and Garcia-Molina.

ETraction3/5

600K page views/day by end of 1998, fully open (no invite). Started with Stanford .edu traffic, spread worldwide via word-of-mouth in six months.

FMoat2/5

Page published the PageRank paper in January 1998; the patent was held by Stanford (later licensed exclusively to Google). The algorithm alone was thin, but the Stanford exclusive license formed a defense line.

GRevenue1/5

Zero. Search OEM licensing to Yahoo! came in 1999; AdWords arrived October 2000. As of 1998, 'how will this make money?' was unanswered.

HFinancial2/5

Bechtolsheim $100K + David Cheriton + Ram Shriram + Jeff Bezos and other angels summed to about $1M, funding 1998. The $25M Sequoia/Kleiner Perkins round came June 1999.

IEfficiency4/5

Four-person team + $1,700/month garage rent + borrowing university compute meant nearly zero marginal cost. Burn multiple was extremely low.

JExit2/5

In 1998 the founders offered to sell to Yahoo! for $1M and were turned down. Search-industry IPO precedent was limited, and in 1998 no exit was articulated.

KTiming4/5

Peak dot-com bubble, 150M web users, AltaVista's quality starting to slip. A perfect inflection point, but the bubble burst (March 2000) was just two years away.

LInsight5/5

'The link graph itself is the ranking signal.' In an era when AltaVista ranked on word frequency and meta tags, and Yahoo! curated by hand, only Page + Brin articulated 'we can mathematically extract the self-organization of the web' (all three v2.0a conditions Yes, validated by Alphabet's $2T+ market cap).

See other cases

View all cases

NGRAPH — AI ADOPTION SUPPORT

We start AI adoption by understanding how your business is built

At NGraph, engineers come into your workplace, look at your work, decisions and the data you use, and then build AI in. We don't stop at teaching you how to use it: we leave AI in a form your team keeps using.