Insights

Questions the corpus can actually answer. Everything below is computed live.

The .com gradient

In 2015 Graham wrote that if you have a US startup called X and don't have x.com, you should probably change your name. His evidence: every one of the top 20 YC companies by valuation owned it, 94% of the top 50, but only 66% of the batch then current.

Today 82.4% of YC's top companies (n=91) sit on their exact-match .com. And the year-on-year decline since he wrote is steep.

Share of each year's cohort whose website sits on the exact-match .com. The dark bar is 2015, the cohort Graham was describing: we measure 66.7% against his stated 66%.

Outcomes

Read this as observation, not causation. It is badly confounded by cohort: companies from earlier years had both more .com availability and more years in which to reach an exit. The table cannot separate those.

Status	n	Avg length	Avg nominology	Owns .com
Public	23	9.6	69.7	78.3%
Acquired	815	8.1	67.5	55.3%
Active	4,292	8.6	57.8	36.3%
Inactive	1,072	8.7	61.6	45.4%

Renames

2,657 companies in the corpus changed their name. Guzman & Stern found growth-oriented startups carry shorter names; Green & Jame found fluency-improving renames precede gains in liquidity and value. Both predict renames should trend shorter — and 55.1% of these did.

1464 shorter

966 longer

227 same length

Two populations

The rubric was written about startups. Running it over the largest US companies by revenue — names mostly chosen decades earlier, under different constraints — is the clearest test of whether it describes naming in general or one culture of it.

Source	n	Avg nominology	Avg length	Owns .com	Commonest shape
Y Combinator	6,202	59.8	8.6	40.5%	invented (34.0%)
Forbes Global 2000	2,000	48.7	13.1	40.2%	phrase (59.2%)
Fortune 1000	1,000	57.7	12.3	66.1%	phrase (54.9%)
Unicorns	836	71.8	8.5	72.3%	invented (45.3%)
TinySeed	205	57.9	9.1	44.4%	invented (35.6%)
Forbes Cloud 100	100	71.6	7.7	75.9%	invented (39.0%)

They barely resemble each other. Fortune 500 names run far longer and are overwhelmingly multi-word phrases, where startups reach for invented single words. Read that as a difference in convention, not in quality — the rubric encodes what one community currently prefers, and scores the other population lower for not sharing it.

Naming by sector

Within Y Combinator, conventions differ sharply by what a company does. Healthcare and Industrials name roughly three characters longer than Fintech, and reach for an existing English word about half as often — deep-tech sectors coin, consumer-facing ones borrow.