Sometime in that first month or two, I heard about a secret project, discussed in hushed tones. And it really was a secret project, at least insofar as such a thing could be accomplished in a tiny office of 100 people on two floors.
The project was so secret that they had forked the code base for the retail site. The site’s web server was called Obidos, named for the town near the fastest part of the Amazon river, and leadership had forked (copied) it and fittingly named it Varzea, after a big swamp. They’d cherry-picked some of the top engineers to work on Varzea, also in total secrecy.
I’ll pause here and get Life Lesson #1 out of the way, which is: Secret projects are stupid. Forking your code base is also stupid.
The whispered secret project? Amazon was gonna kill eBay. The plan was to build an entire auctions site from scratch, fill it with some sweet seed content like rare collector baseball cards from Sotheby’s, offer an ironclad guarantee (“We’ll refund you if our sellers screw up”), and then pounce. Boom, bam, the tab launches, the eBay sellers all pack up their shops and leave eBay en masse, all the eBay customers stampede over as well, and eBay is left sitting there wondering what hit them.
Great plan, eh? Well, of course we can laugh in hindsight, especially after a few drinks. Honestly, even at the time it should have been obvious that it would never work, because it’s not as if eBay was the world’s first network effect.
But Jeff had never been wrong before. And nobody had truly believed that his little bookstore was going to be so big, until it grew like a Cat-5 storm and proved everyone wrong. In early 1995, everyone from my tiny graduating CS class at the University of Washington was looking for jobs at big tech companies like DEC, IBM, and upstart Microsoft. One of our classmates told us she was going to work for this internet bookstore, and I remember feeling so sad for her. We felt like she had decided to throw away all her fancy training and become a librarian or a nun. She was employee #11.
By 1998, post-IPO, Jeff had cleared up any doubts as to whether his bookstore would succeed. Every time he tried something crazy-ambitious, he wound up being right. So when Jeff said they were going to kill eBay (this was before the U.S. government sued Microsoft and everyone learned not to say you’re killing your competitors), people believed him. They just straight up had faith. Amazon had a cult-like culture at the time, a speak-no-dissent mentality, where discussion of failure simply wasn’t allowed—perhaps because people superstitiously believed saying it might make it come true. Google+ was like that, too.
I don’t know why Amazon Auctions had a mascot named Mr. Tooth. My hunch is that Jeff came up with it himself. There’s really nothing wrong with Mr. Tooth as a mascot. He wasn’t a bad idea. The bad idea was Amazon Auctions — because you can’t beat a network effect with a nearly identical network.
Amazon Auctions wasn’t fundamentally different from eBay. Sure, it looked a little cleaner, which is unsurprising given that eBay’s site always looked like someone threw up all over it. And Amazon had established enough brand trust that the refund guarantee was legit. I have no doubt everyone believed it.
But Amazon Auctions was a direct competitor to eBay, which meant that in order for it to be successful, people were going to have to leave eBay.
And that, dear friends, did not happen. Amazon Auctions was a spectacular flop. Put in modern terms, it was about as successful as Google+. And for the same exact reason! The problem is that certain ecosystems exhibit what’s called a “network effect,” in which the system naturally reinforces itself through a feedback loop, causing an almost magnetic attraction back into the system.
Even the best marketing isn’t strong enough to beat an entrenched network effect.
In eBay’s network, the buyers go where the sellers are (for variety), and the sellers go where the buyers are (for reach), and it becomes self-reinforcing. Buyers are busy people, so they would visit Amazon’s Auctions site, not find what they wanted, and head straight back to eBay. Sellers have more time on their hands, and some of them tried Amazon Auctions. But they weren’t getting bites, so they started decorating their sites to point back to eBay, where all the buyers were.
In practice, the only way to get everyone moved over to an identical network would be to do it all at once, and of course that’s logistically impossible.
So, Life Lesson #2 (and this one is pretty goddamn important) is: Don’t try to beat a network by making a clone with improvements. It ain’t gonna work. There is too much gravitational inertia in the original network; nobody is incentivized to leave it.
There are some approaches that will work, and we will explore two of them: Jeff’s and Jack’s. Both require you to get into an entirely different market and build a network there. You can’t beat a strong network head on, but you can flank it.