The Brutal Logic of Survivor Bias
One of the humbling privileges of the job that I have as a venture capitalist is that I get to talk to so many amazing, committed, smart, driven, talented entrepreneurs, builders, designers, operators, marketers, sellers, and on and on. And I get to talk about so many different aspects of what they’re doing, in such detail - it’s really an unbelievable privilege, and I’m grateful.
But one of the things that comes out of that is that sometimes I visit with entrepreneurs who are struggling to make things work, who are struggling to sell, or get attention, or get product market fit. And like the world-changing alphas they are, they wonder “WTF? Why isn’t this working? What’s wrong with me?”
And the truth is this: starting something from nothing is *hard*. Nothing comes easy. Competition is brutal and global and 24/7 unrelenting. Inventing is never obvious.
But we have a survivor bias problem. The companies and people we all compare ourselves to are the ones who won, and who are winning. The Googles of the world. The Dropboxes. Ev & Biz & Jack. We compare ourselves to some of the most outrageous successes in the history of the world. We compare ourselves to the winners and the survivors.
That’s a hard standard to live up to for anyone.
It’s useful, of course, because it helps us dream bigger. It helps us think more about what’s possible. More “why not us?” And that’s great.
But it’s tough to see the forest for the trees sometimes and remember that for many, it wasn’t an immediate success. For many, they’ve tried before. For many all the other, less successful endeavors in their lives turned out to be leading them to their ultimate success.
So that’s the message I give, really. Don’t sweat the hero story so much. Strive for it, but focus on the business and product and team you’re building. Remember that the company may have more chapters yet to be written, and that the people in it most definitely do.
Survivor bias creates a tough standard when it looks like everyone is “killing it” and “up and to the right” - but don’t let it distract you from doing the thing that *you* want to do, and don’t forget that sometimes we struggle to a breakthrough.
Facebook’s new Paper app looks really, really lovely. It’s interesting that one of the lead value propositions they espouse is that it’s “ad-free.” But if they really continue down the path of unbundling their apps (messenger, updates, etc), they start to lose some of the advertising impression power they get now from their integrated feed app. So if Paper works — if people really adopt it, feels inevitable that over time it won’t be ad-free. (Which, honestly, doesn’t bother me personally in the slightest, but YMMV.) But maybe it’ll allow better feeling, more targeted ads. Might be that charging the publications (they show Time & The Atlantic) a distribution fee might work.
In any event, looks like some really, really nice work by the product team there, and I’ll be excited to try it out on Monday.
If you care about the NSA surveillance programs and the state of our republic (and I submit that you should), go read at least pages 42-56 of this ruling in Klayman v Obama by Federal Judge Leon in the District of Columbia — the high line is that the program most likely violates the Constitution (ed: shocking!), but there’s a lot of very interesting stuff in his ruling. Here are a few choice quotes:
"Indeed, the question in this case can more properly be styled as follows: When do present-day evolutions in the GoVernment’s surveillance capabilities, citizens’ phone habits, and the relationship between the NSA and telecom companies become so thoroughly unlike those considered by the Supreme Court thirty—four years ago that a precedent like Smith simply does not apply? The answer, unfortunately for the Government, is now."
"To the contrary, for the following reasons, I believe that bulk telephony metadata collection and analysis almost certainly does violate a reasonable expectation of privacy."
"It’s one thing to say that people expect phone companies to occasionally provide information to law enforcement; it is quite another to suggest that our citizens expect all phone companies to operate what is effectively a joint intelligence- gathering operation with the Government."
"…the almost—Orwellian technology that enables the Government to store and analyze the phone metadata of every telephone user in the United States is unlike anything that could have been conceived in 1979. In Smith, the Supreme Court was actually considering whether local police could collect one person’s phone records for calls made after the pen register was installed and for the limited purpose of a small-scale investigation of harassing phone calls. See Smith, 442 U.S. at 737. The notion that the Government could collect similar data on hundreds of millions of people and retain that data for a five-year period, updating it with new data every day in perpetuity, was at best, in 1979, the stuff of science fiction. By comparison, the Government has at its disposal today the most advanced twenty-first century tools, allowing it to "store such records and efficiently mine them for information years into the future."
"…people make calls and send text messages now that they would not (really, could not) have made or sent back when Smith was decided—for example, every phone call today between two people trying to locate one another in a public place. See CTIA Quick Facts, supra (2.3 trillion voice minutes used in 2012, up from 62.9 billion in 1997). This rapid and monumental shift towards a cell phone-centric culture means that the metadata from each person’s phone "reflects a wealth of detail about her familial, political, professional, religious, and sexual associations," Jones, 132 S. Ct. at 955 (Sotomayor, .,concurring), that could not have been gleaned from a data collection in 1979."
"Rather, the question that I will ultimately have to answer when I reach the merits of this case someday is whether people have a reasonable expectation of privacy that is violated when the Government, without any basis whatsoever to suspect them of any wrongdoing, collects and stores for five years their telephony metadata for purposes of subjecting it to high—tech querying and analysis without any case-by-case judicial approval. For the many reasons set forth above, it is significantly likely that on that day, I will answer that question in plaintiffs’ favor."