So having a founder with past experience in a successful company gives you a slightly lower chance of success than taking a random first-timer. The existence of wildly successful second-timers does not answer this fact. This slight anti-correlation predicts the existence of wildly successful second-timers. Just fewer of them than you'd get if you selected an equal number of random first-timers.
Secondly, comparison to an existing major player is never a good move for a startup. One of the more frequent arguments I've had in startup environments has been countering the claim that "Microsoft (or Apple, or whoever) does it this way!" as justification for some particular business practice. It is surprisingly difficult to gently disabuse a sincerely committed individual of the notion that their ten-person zero-revenue startup is meaningfully comparable to pick-your-multi-billion-dollar behemoth with tens of thousands of employees, a universally recognized brand, a world-wide distribution network, and a generation of technical expertise.
So there is in fact little more than theory to support the validation. Factual validation necessarily involves paying customers. Lots of them.
'Til then, this looks like a late entry into a market dominated by Amazon, with Amazon Prime already owning the membership space for online sales. Unless their execution is perfect in every detail (which is could be but more likely is not) and/or they offer some compelling advantage over Amazon's delivery model and Amazon Prime's membership model it's hard to see how they are going to put significant dents in Amazon's position. Maybe they don't have to, if they're lean enough and clever enough, but that's a pretty risky bet.
costco (and even amazon prime) prove the membership model
past experience proves the founder can build a large ecommerce company
there is a little more validation than just theory