That's missing the point of databases though. Do you really want to manage all your orders via emails?
I've done a low tech store. Basically users browse products they like. They enter their phone and address to send it to as proof that they're actual buyers. The page caches their contact details so they can do the same on other items. If it's hard to imagine, it's in the cover image of this article: https://smuzani.medium.com/setting-up-a-mobile-backend-serve...
It goes into a database. We'd filter orders by phone number to form a "cart", and also filter out "delivered" statuses from their past order.
It was meant to be a MVP to see if people in one country would actually buy ingredients through a recipe app. Worked incredibly well - we were swamped with orders as soon as we allowed it.
It was around 85 items over 3 days, totalling around $750. We'd contact them via WhatsApp or SMS, request payment, verify payment, send the item, then verify it arrived. But we had the power of databases to sort that out. That stuff gets exponentially more complex the more orders and items you get.
But why would you avoid doing that to save on "database costs"? It's a lot harder when you can't really track who you contacted, who paid, who said they were going to pay but haven't paid, and so on. I'd have happily outsourced that process for a few cents.