First-Best Premium, Second-Best Discount
Against the first-best bias
Welcome to the eighth edition of Black Box. Last time, I synthesized the four articles I have written so far into a single thesis about culture. Most articles going forward will fall under one of its four themes. This time, we revisit Culture as Communication.
What makes working at Cameo interesting to me is the fact that we’re one of the few companies that contributes to and is influenced by pop culture. Our Slack is filled with crazy stories of Cameos in the wild that are impossible with other products. But there are also unique challenges to being so involved in culture. One such problem is the belief that Cameo only has B-list celebrities, so anyone joining Cameo must be past their prime. This idea is baseless—last I checked, Lil Nas X and Snoop Dogg are still hugely popular — but the stigma is enough to make some celebrities wary of signing up. Understandably, no one wants to be second-best.
But the stigma around second-best is itself baseless. Second-best is not just as good as first-best — it is often better in terms of return on investment. I believe second-best talent represents the most valuable and under-tapped cohort of human capital across industries. Hiring managers, investors, marketers, and other decision-makers should recognize this arbitrage opportunity and take advantage.
Convenience as Original Sin
Before I continue, it is important to understand the difference between first-best and second-best. An entity is first-best if it is what most people think of when they think of its category. New York, Harvard, and Goldman Sachs are all examples of first-best entities. It doesn’t matter that San Francisco, Cornell, and Morgan Stanley are equal or better in many aspects because first-best is rooted in culture, not substance.
This raises the question of why there is a distinction in the first place. After all, second-best entities have significant merit of their own. I am not a sociologist, but I think some combination of the following explains this division:
Elite overproduction is the idea that society produces more potentially elite members than it can absorb into its power structure. One solution is stratify the elite class further, thus creating a need for a first-best.
A Schelling point is a choice that people who are attempting to coordinate default to when they lack information about each other. First-best entities make good Schelling point references because they are widely known.
Synecdoche is a literary device that uses a part to refer to the whole. As synecdoche, first-best entities are abbreviations of their categories.
In taxonomy, a type species is a species whose name also identifies its genus. Type species are logically similar to synecdoche except they are unique and institutionalized. This is the case for many first-best entities.
The common theme in these ideas is first-best entities provide convenience. Their existence enables people to skirt work by reducing the complexity of their problems and considerations. In time, first-best becomes the preferred solution or conclusion — even if it is not the best or the only.
Gold Plating and Brass Tacks
Perhaps no category has a preference for first-best as strong as educational credentials. Investment banks and consulting firms famously recruit at only two dozen or so target schools, but they are in fact on the lenient side when it comes to alma mater. Big Law obsesses over the T14. 40% of VCs graduated specifically from Stanford or Harvard. I was surprised to learn that journalism may be the worst offender with Columbia, but this made sense once I realized its graduate school is literally home to the Pulitzer Prize.
Do these schools make that much of a difference? Ironically, I asked this very question for my economics thesis when I was at one of these institutions. My research turned up a research paper that looked at exam schools in New York and Boston (e.g., Stuyvesant and Bostin Latin). The paper used the admission cutoff in these schools’ entrance tests to perform a type of natural experiment analysis called a regression discontinuity. The idea of RD is given a threshold, people just above and just below it should be essentially the same prior to the treatment — in this case, attending a prestigious high school. Any differences after matriculation can therefore be attributed to the school alone.
But there weren’t any differences. These two groups of were indistinguishable statistically in standardized test scores (state assessments, AP, PSAT, and SAT) and college outcomes. This was despite huge differences in the reputation and socioeconomic makeup of exam schools versus non-exam schools. Indeed, the authors call out that “while the exam school students in our samples typically have good outcomes, most of these students would likely have done [equally] well without the benefit of an exam school education”.
The implication for hiring managers is that they should expand their pedigree horizons. Doing so not only has no impact on candidate quality, but also helps them avoid the first-best premium. After all, even non-econ majors know that lower supply means higher prices (all else equal). This effect is also familiar to investors. I cited this paper as an analog to the actual topic of my thesis, which asked whether the high valuations of Y Combinator startups is justified. Based on funding distribution, follow-on funding rates, and exit size and probability, my conclusion is similarly no.1
Playing Tennis with Micro-Influencers
If first-best premiums are unjustified, how much more value is produced with second-best talent? Quite a lot, as it turns out, because of a phenomenon I call the second-best discount. A stark example is professional tennis.
Pro tennis players are ranked according to the ATP and WTA systems for men and women, respectively. Players are awarded points based on the prestige of the competition and their performance, but their rank is based on only the top results due to a limit on how many games can count. The highest-rank players make incredible amounts of money thanks mostly to endorsements — Federer became a billionaire last year. Yet younger players who are just outside of the top 100 struggle to break even. The New York Times reported that №139 Chris O’Connell made just $15,000 in 2019, despite winning 82 matches that year. In other words, one of the best tennis players in the entire world was effectively earning minimum wage.2
What makes this disparity so maddening is that the lack of resources actively holds back players on the cusp from breaking into first-best. Full-time coaches are critical to improving, but most second-best players can’t afford one. They train in subpar facilities and eat instant ramen in cheap motels when they are on the road. Some second-best players literally can’t afford to travel to enough tournaments to advance, given the cap. Second-best becomes a self-fulfilling prophecy despite their talent and potential because their value rounds to zero amidst the hype around the first-best.3
Influencers and creators also face extreme power law dynamics. While Charlie D’Amelio made $17.5 million last year, nearly half of full-time creators earned less than $1,000. This is a huge opportunity for digital marketers because that half includes the second-best creators commonly known as micro-influencers. Although these creators have smaller audiences, their followings are stronger and more engaged within their niche. This is particularly valuable with iOS 14 privacy protections making targeting more difficult for marketers. Combined with their far cheaper price, micro-influencers produce much higher ROAS than internet celebrities despite the apparent difference in fame.4
What’s interesting about the divide between first-best and second-best is these micro-influencers would get additional sponsorships if more marketers “took advantage” of this arbitrage opportunity. Talent from a wider range of schools would have a chance to break into elite industries, and up-and-coming tennis players could finish coming up and carry the future of the sport. Second-best people need a shot, and decision-makers can give them one by thinking more about return on investment and less about pedigrees and vanity metrics. Only then will the gap between first-best and second-best truly close. ∎
Have you seen a first-best premium or experienced a second-best discount? Share your thoughts @jwang_18. See you in two weeks!
There may even be a first-best premium at the intersection of schools and startups. I have noticed that founders of well-funded startups outside of the U.S. tend to have studied at elite U.S. business schools — see Coupang, Elenas, and Mercado Libre. (Then again, this could be due to the networks that they developed there. Does anyone have data on either hypothesis?)
Working 40 hours a week for 52 weeks at the federal minimum wage of $7.25 per hour yields an annual income of $15,080.