Sean Ellis

But prior to that, he had a company called PhoneTag, where he had basically done a ton of testing and came up with this double-sided incentive as being really effective.

But PayPal did the same thing.

And so in our case, the currency was storage.

So it just made sense.

So it was really something that...

three of us said we should do this and basically pulling the pieces together of how it was done was again an iterative group effort.

Yeah, I think the first part is, again, like I think in the case of Dropbox, people credit the referral program as being something that drove a lot of the growth.

Referrals were strong before the referral program went in place.

People love the product.

We're spreading the word about the product.

So I think you can't band-aid kind of a crappy product experience and a crappy product value with a referral program, for example.

So that's the first thing is I just make sure it's happening.

And then understand what the motivation is for people to do it.

And then think about really every step in the process from what's the prompt that gets someone to share to what's the invitation or the person on the other end seeing.

AB tests a ton of variations of that.

When they click through, what's the page that they sign up on?

How do you give feedback to the person who sent the initial invitation that what they did worked?

And each one of those steps is something that can be optimized.

And just like any kind of conversion rate optimization, the more you have feedback,

qualitative and quantitative insights about what's happening, you're going to be more informed in the tests that you're running.