I really liked the part about everything scaling. I hate when people say, "It wouldn't have scaled." It would have. You just weren't thinking of it in the right way.
This is solid advice: "You need to find the balance between what you're giving away and the barrier to getting it." It's a common mistake to overlook this.
Honestly, I think it's a given that web personalization is deeply rooted in A/B testing. But instead of testing for what's best for the aggregate, you're testing for what's best for each individual.
That's interesting, but I suspect most people want to maximize visits, in which case they're after the aggregate. I'm not saying I agree with the sentiment, but I think it's true.
This reminds me of the saying: "Never compare your beginning to someone else's middle." It's never too late to start, so it's never too late to be awesome.