This fee structure (3-5%) will not work. Exactly why does Zinc need $50 to process my HDTV order but only $0.15 for my banana slicer?
Funding has no transaction fees due to Dwolla transfers and Zinc is most likely taking a cut from Amazon Affiliate program as well (which pays 5-15% commissions depending on category)
Maybe I'm just having a tough time figuring out the use case for your free tier, but if I make less than 10 orders/month, exactly why would I go to the trouble of implementing an API?
I'm guessing people that are just tinkering with the API for the fun of automation. For example, I could use this to order new socks every few months. Just because, yay, unexpected socks in the mail.
It might be the case that people are just tinkering with this service for fun, but the "yay, unexpected socks" use case seems like a solved problem with Amazon Subscribe and Save lists (http://www.amazon.com/Subscribe/b/?_encoding=UTF8&camp=1789&...). To each their own, but I just don't see myself reinventing the wheel with this API (not to mention paying an additional 5%) to buy socks/paper towels, when Amazon has already done that leg work for me.
I do, however, agree with meritt's sentiment above...zinc.io could be used to create some valuable services for existing businesses (rewards programs, et al), but as a user of this API, I think it'd be much more cost effective as a fixed per-order fee structure.
As Amazon doesn't offer an API for this, how's it done? Scraping? Hordes of low-wage folks filing orders manually? What's stopping Amazon from blocking this?
We use mostly automated processes, with manual intervention only for rare edge cases. One thing that's surprised us is just how many (i.e., non-rare) edge cases there are when you take into account shipping exceptions, order constraints, etc. We believe we are on solid footing both legally and technically, and this has been supported by three years of reliable operation.
Unfortunately we're unable to use most affiliate programs. At the same time, we're mostly interested in building a framework that easily scales across to all merchants.
Why would they need one? Amazon (in terms of the shopping site) is geared to consumers. Even if she could, why would my mother write a script to order things from Amazon? By using the site, you are exposed to their attempts to convince you to buy other products, or a more expensive version of their current product.
And probably much more importantly you're exposed to their data mining. They get to monitor what products you buy directly so they can help refine their recommendations.
You can sell stuff on Amazon. You can have your orders be filled by Amazon. You can display your products for sale on Amazon on your web page. You can use Amazon to charge people money. The only thing preventing Amazon from being a complete white label web store is the fact that you can't hook it all together and run your own store front where people come to your site, and buy something without going through Amazon's site.
I don't assume anything about your mother, but if she ran a business that (1) had some sort of computerized record-keeping and (2) used Amazon-appropriate supplies at a rate that was fairly predictable from those records (although not necessarily predictable per time period), a script targeting this service could save her a lot of time.
Now, Amazon might feel like they don't make enough money from such an unimaginative customer to justify the cost of the service. I don't see how it would steal sales from them, however, until it starts using other suppliers as well. Maybe they'll acquire before that gets turned on?
You'd think, but you'd often be wrong. If you're a small retailer, even with the "wholesale discount" ordering from the major US wholesalers, you'll still be undercut by Amazon fairly often. You don't get the equivalent discounts until you're ordering tens of thousands per order from individual suppliers, if ever. That means you're not just a busy store but a busy chain of stores.
My parents own a health food store. They pick up a few of the products they carry whenever they're at Wal-Mart, because Wal-Mart's retail is below the manufacturer's wholesale rate for them. Obviously they don't make much margin on those things, but their customers want them to carry them anyway.
Amazon is fine with acting as a supplier for other stores. They will take sales tax exemption forms for reselling their items.
Maybe, but a common API that can order millions of different products with reliable delivery/billing/returns seems like it would be very useful. Dealing with middleman distributors (who often want you to fax orders in) is a nightmare for many businesses.
Amazon offers better prices than many wholesalers, and much better service than most, in that lead times are shorter and more predictable. If you guarantee your customers you can always install 10 widgets, and you keep 30 in stock, what happens the day after you install 10 widgets for three different customers? With this sort of setup, you receive another 30 widgets by FedEx. Let Amazon do the floor-planning, and spend those resources on something else.
Of course, there's little to prevent this service from signing up big wholesalers as well, if they're able to meet its service requirements.
If done manually, it would also a great example of identifying a need before building a product. As long as it looks automatic why would the API user care.
Thanks to everyone for your comments and feedback. As we shared on our site, we've been automating ecommerce tasks for a while, but this is our first attempt to open our tools up to the public--and we're learning.
We recognize that our current pricing might not make us suitable for all uses, but we think it will work for enough applications to support us as we build out a killer product.
We really appreciate your feedback as we work to improve what we view as a very flawed ecommerce environment.
We can't share all the specifics of what integrators are building, but some are very clever and surprising. Some are automating business processes and other are creating niche resale stores, but the most fun are the creative/novelty uses we've seen.
How exactly do I identify the product I want to purchase? There can be many slightly distinct versions of the same product on amazon. It sounds like I would have to pre-select the products that my users would be purchasing, because for them to search for what they want they'd have to use the amazon website anyway.
I'm just having trouble seeing the value in this, except for the very specific use case of regular, event-driven, identical, automatic orders (eg, I have my lights hooked up to an intelligent system that notices that the bulb blows out, so it then submits the pre-prepared order for the correct bulb via zinc.io).
Are the prices really so bad? Those fees seem comparable to a payment service provider. This is a different service with presumably lower risks, but if you have to buy stuff and this service replaces a significant portion of an FTE it could definitely be a "reasonable cost of doing business".
I was thinking the same thing. Why isn't this a free service that collects an affiliate fee instead. Seems like it would lower the barrier for adoption.
The end-user is not clicking through a referral link, going to the Amazon site and making a purchase. How would Zinc earn the affiliate fee? If you're suggesting they negotiate a custom referral arrangement with Amazon, that's probably impossible at the MVP stage.
They might be filling orders manually as some suggested. A Turker working for Zinc.io and placing the orders via their affiliate links would do the trick. No need for negotiations etc.
You can't refer yourself, else the program wouldn't work -- everyone would sign up, everyone would click their own links, and Amazon would effectively have dropped its prices across the board while adding a rebate-processing overhead to every order. That's why program agreements virtually always prohibit this.
Stripe is a hit with developers because of their easy API and great documentation. It's not a surprise that other "by techies for techies" API startups are taking inspiration from a leader in the industry.
Why would you send "A Million Random Digits with 100,000 Normal Deviates" to Norman Borlaug, the inventor of Dwarf Wheat? Is this some kind of hipster pranking?
But what's stopping them from collecting the affiliate fee?
This narrows the use-cases for this quite a bit. Not only are they forcing you to forgo collecting the fee yourself for somebody you referred to Amazon to make a purchase (assuming this is a service/app you've created), you have to pay _more_? So really, depending on the model, you'd have to mark up the Amazon product ~10-13% just to cut even.
Amazon Associate's TOS. You can't get affiliate fees for products you purchase (technically they are still the purchasing agent--you even contact Zinc for a return label).
Funding has no transaction fees due to Dwolla transfers and Zinc is most likely taking a cut from Amazon Affiliate program as well (which pays 5-15% commissions depending on category)
Make it a fixed per-order fee.