Thoughts on the "flywheel" (and other ideas from the book "The Everything Store")

Very interesting take. I agree that the third-party sellers were a huge key in getting their flywheel turning:

better customer experience → more traffic → more sellers → more products → repeat

Allowing third-party sellers to work on commission obviously represented a boost to the "more sellers" part of the cycle. I also agree that this was a non-obvious decision. However, I'm not convinced that being non-obvious is what made it effective.

Rather, I think the flywheel's spinning will always be hindered by its weakest link, and the arrow between more traffic and more sellers was probably the weakest link for Amazon.

When making a flywheel for IH, I realized something interesting: not only should you measure the components of the flywheel, but it's also crucial to measure the arrows between the components. If the arrows aren't doing well, it means improvements in one component won't lead to improvements in another, so the flywheel won't turn.

In Amazon's case, more traffic is obviously enticing to all sellers. However, there's a lot of lag and a lot of friction there. How do sellers know Amazon gets more traffic? How quickly can they react to small changes in traffic? Do they even react at all? If Amazon has to call up sellers on the phone and say, "Hey, our traffic increased by 5% in your category, let's renegotiate our contract," that's a lot of friction that stops the flywheel from turning.

In other words, Amazon might improve its traffic, but that doesn't mean much if the arrow between "more traffic" and "more sellers" is weak.

I think allowing third-party sellers in the door was a huge break there. It was like adding liquidity to the stock market by going from 100 traders to 1,000,000 traders. Suddenly sellers could react instantly to increases in traffic, increase their inventory, and tell their colleagues. "More traffic" could now automatically lead to "more sellers" and "more products" without much friction.

For IH, biggest point of friction is the arrow between happier users and more traffic. I measure this by looking at retention and the number of repeat visits per week. The rest of the flywheel turns itself pretty well — more users leads to more threads, and more threads leads to better threads at the top, which people read. But things drop off when it comes to retention, so that's the area I need to focus on.

I also have another part of my flywheel, an outer loop, that goes like this:

more threads → bigger variety of threads → happier users who return more often

It's a bit unintuitive unless you read Hooked or study how Facebook works its magic, but the basic gist of it is that people get bored of seeing the same old stuff all the time. They won't make a habit to visit something repeatedly unless it's new every time, like the HN front page. So it's crucial for me to get to the point where 50-100 new threads are created every day.

I think your capital injection idea is interesting. One advantage we have with IH is that we already talk to high profile guests via the podcast and text interviews. I'd like to do more to get them involved in the forum, as they have considerable audiences they can bring, and they also add a lot of clout, recognition, and valuable knowledge that can help make the forum a worthwhile place to return to.