Thank you for the thorough and thoughtful reply. I’ll reply to your points about scale and federation first.
You’re right, I think the grouping OFN refers to is within an instance. Certainly I’d expect the W3C group to know if there was any significant federated-shopfront software already out there. I don’t know if ActivityPub might be a good basis for a federated shopfront (or shopfront plugin).
I’d agree that a large store would probably not be a use case for FreedomBox. I was also thinking more of the average-Etsy-scale store, with single- or double-digit sales per week. Perhaps it’s better seen as two separate problems:
- A small merchant wants to sell widgets online (say, a crafter who might otherwise use Etsy, or an independent bricks-and-mortar store that wants to list their products and prices online).
- A purchaser wants to buy a widget (they can probably find it on Amazon, or in a big box store, but they are not going to check half-a-dozen independent shops).
2a. A merchant saying “Why do they buy widgets at [Amazon/franchise competitor] when my widgets are cheaper and better?”
I think discoverability is the only motive for federation; a cross-store search function is valuable to both parties. I’d love to be able to search by product barcode or keywords and find out who stocks what where at what price. There have been many commercial providers of price-comparison sites, so I’m not the only one. Unfortunately, the business model for most price-comparison sites tends to degenerate into manipulating results according to “advertising fees” paid by merchants (e.g. Google Shopping). Walled-garden aside, these conflicts of interest make the comparison sites pretty useless.
Discoverability is not a big deal in some markets. If customers can find all the local stores by driving down a couple of streets, or most customers have lived in the same place all their life and talk to each other a lot, they are likely to know about all the local stores for products they buy regularly. For selling and buying food in an Indian village, I think a local e-commerce portal shared by multiple local shops would work very well. I’ve even seen a 20th-century offline bricks-and-mortar version of this model, and it worked.
If customers are savvy searchers and willing to spend a lot of time hunting for a product, discoverability is also less important. If the product is frequently talked about in community of potential customers, word-of-mouth can do wonders; you can even self-host a crowdfunding campaign. Some leaving-the-cloud-type products, hardware and software, have successfully relied on this. These are intrinsically unlocalized products, due to the economies of scale for electronics and software (and customizability).
But if you are selling a useful but boring product to a small, non-localized market of socially-isolated, footloose, busy customers, discoverability could be your biggest worry. This is a common situation, faced by anyone trying to start or run a specialized store, or a store in a megacity. An independent website selling your typical Etsy/Amazon product might genuinely get less turnover than an equivalent Etsy/Amazon storefront.
Since having more sales through your [Etsy, Amazon etc.] storefront boosts your visibility, you have a strong disincentive against making sales off-platform, even to your own mother. 
So you pay the 5% commission on each sale as an advertising fee.
I’ve talked to some of these types of storekeepers, and I think non-federated e-commerce software in FreedomBox would be useful for many. A federatable system would be still more useful for some people, but not all. Wordpress plugins like Woocommerce sound useful.
