I think it's easier to take a traffic source and build a product around that, than to build a product and then find traffic.

An idea for a traffic source is a gimmick. i.e. something you build that brings people to your site.

The best example I've seen is https://blockchain.info now called blockchain.com/explorer.

They had a site that gave a web based view of bitcoin addresses, balances and transactions.

This gained a huge amount of traffic in the early days of Crypto (2012).

Then the rest of the products, i.e. wallet and exchanges came later and sit on top of that traffic.

So my advice to myself. Traffic first, then build a product.

If the product fails at least you learned something and still have the traffic.