Comment by an_taeus

Outstanding! I worked for a number of years at one of the larger used book resellers in the USA, top 1-3 of Amazon's used listings consistently, it had a fairly surprising amount of volume to me when I first started. The fact that Amazon was 98%+ of our sales source was considered an existential threat to the company everyday but the ability for us to mitigate that was almost impossible. Even after multiple temporary pauses on our Seller Central account (because of bogus customer complaints sent from our competitors/buyers who were buying our used inventory, then re-listing it on Amazon as "new" at a higher price point, and who then repeatedly gamed Amazon's refund policy to pay nothing for it and then "return" the items, even though we'd stamped them and knew it was bogus and hadn't come from us...) the owner never dedicated the resources it would have required for us to truly become multi-channel. The volume from Amazon, especially for books, is ludicrously profitable compared to any of the other e-marketplaces but terrifying from a risk and sustainability POV. This is a major pain point for so many small businesses out there and I wish I had this while I was still there, I will be forwarding it along to those who still are. Now that I'm thinking about it, the hardest part for us with Amazon was that we historically were treating our Amazon listings as our inventory counts, and let's just say Amazon doesn't do a great job with handling items that are clicked to buy but cancelled afterwards. There's absolutely no way that the company I worked for is the only one that does or did track its inventory by coupling itself to the sales channel, so you've probably saved a lot of people from future headaches already.

One thing I think you've got very right here is the relatively spare use of fields in your schema models. I can't emphasize enough how much that will help a seller that uses this, since so many small business owners and employees out there don't know what they don't know and are making it up on the fly and tend to focus on the wrong things. Not a lot of headspace for the "nothing missing, but nothing unnecessary" approach when you're trying to juggle every hat because you're understaffed, and doubly so when you've never even heard the term many-to-one. Of course, because this is open-source, if users want they can extend these schemas (and even extend the views to use them! etc.), but they'll have to go out of their way to do that, not start with a tangled mess of concepts and pare them down. Really, it seems like no big deal, but humble schemas are so so important. At my current job I'm having the fun time of maintaining true spaghetti software with hundreds of fields in tables, a lot of which are null (or empty string! but which is it?), and all of which have no field names beyond a number...to power a POS/inventory system. But don't worry, none of the data is normalized and it pre-aggregates your transactions to compress itself, and then writes those compressed records into an always existing "temp" table and then reads them every time you want a report, and look at that, it takes 5 minutes to generate a COGS report for last week that is a sub 5 ms SQL query. Ugh.

I guess where I'm going with this is that there is so much crapware out in the world, absolute trash that needlessly overcomplicates the problem it addresses, that continues to get its license renewed every year because the clients its sold to now believe its a truly hard problem, but if you look under the hood it's actually just some WSYWIG editor generated garbage that is trying to be everything and so instead is nothing usable, and their sales staff will tell you "just upgrade to our new microservices design to solve those legacy issues!" Every project like yours that takes what is a fairly traditional, well-trod problem and brings it into the open-source realm in a readable and straightforward way is a big win for small-time entrepreneurs. Please keep going with this.

I think you have something here, but the hardest part will be marketing it if you're trying to do this for a living. There are absolutely an absurd number of small businesses out there living or dying by their ecommerce sales who are either struggling to actually use the tools and workflows they currently have or are paying far more than they need to for sub-optimal solutions to this problem. They will not come to you first over the internet because they are not looking for open-source sales software, because they don't know what open-source is, but some honest shoe-leather work pounding the pavement and identifying the larger businesses around you physically or trawling through the business formation records in your county might pay real dividends, even if it's just an HN "lifestyle" business for you with recurring revenue comparable to the average salary.

As a final thought, when you do get to integrating Seller Central, you should look at bringing the different API services out there into your project for dashboards and the like. Maybe you don't want to deal with all that, but being able to flip through market metrics from all the different sales channels and even integrate them together would be a killer feature. In particular I'm thinking of https://keepa.com/#!, but I know other people use https://www.rainforestapi.com/ too. Don't know about the other e-commerce marketplaces out there but I'm sure they have them too.

Congrats on a getting a ten year long lurker to actually post.