Here's my take on it and someone can chime in if I'm wrong on the details...
Imagine you're a startup/inventor that manufactures a widget that you'd like to sell to the mass market, you could either (1) try to sell direct from your website (2) get on the shelves of national stores like Best Buy, Target, Walmart, etc. The "shelves" include the real ones at their brick & mortar stores and the virtual shelves on their online websites.
(1) is hard to attract shoppers since nobody knows about your low-traffic website. Also, you'd have to handle the hassle of fulfillment yourself. Amazon Launchpad leverages their competency in global logistics to do this for you.
(2) is difficult to get meetings with corporate retail buyers and convince them to carry your product. Sometimes, there are also "slotting fees" (sometimes aka "bribes") to get prime shelf locations (eye level vs the floor.)
What Amazon is doing is opening up their "shelves" which includes the prime pixels real estate on their front page to promote startups' products. They are actively marketing your product. This is a different initiative from passively showing the 3rd-party marketplace sellers on amazon product pages.
However, to filter out the low quality junk and avoid every garage warrior trying to sell their flavor of homemade barbeque sauce, the products have to come from "the approved network"[1] that includes firms such as a16z, Accel Partners, etc. If we scan that list of affiliates, we'd expect the products vetted by them to be "cutting edge" and "innovative".
What's not specified in all the press releases and FAQ about Amazon Launchpad is the type of payment structure Amazon expects. Is it negiotiated on a product-by-product basis? Is it flat percentage?
> However, to filter out the low quality junk and avoid every garage warrior trying to sell their flavor of homemade barbeque sauce, the products have to come from "the approved network"[1] that includes firms such as a16z, Accel Partners, etc.
In other words, Amazon will give preferential advertising space to those well-connected with the Silicon Valley elite.
PG's heuristic for this is 'write like you are explaining to a friend.'
The problem is that that kind of clarity is rarely achieved on a corporate website because Conway's Law applies as much to marketing as it does to software design.
This is what the NLP community would call 'abstractive summary'. It fundamentally requires generating new sentences - which is an excruciatingly impossible task for current AI tech, at least as far as my understanding goes. The closest tech is the sequence-to-sequence deep networks. I am yet to dig deeper into this subject.
There was recently a Google result here[1] where they were able to summarise a paragraph into one sentence. It looked pretty convincing. How does your "excruciatingly impossible" stand when confronted with that? Were the paragraphs in that result particularly easy to summarise? Does the problem's difficulty grow super-linearly with length of the text? Is there something else in the way?
I would love to see that on some actual text (rather than published "samples"). I expect the more text it has to deal with the more unintelligible the result.
That's intuitive. What isn't is how much harder it becomes with longer text. Imagine there is an abstract measure of "smartness" of your algorithm (that would intuitively have some relations to your model size, teaching sample size and so on). My question could be worded as "what is the asymptotic smartness complexity of the summarisation problem?"
Btw, I'm not aware of existence of such a measure, which I take as a signal it's infeasible to formulate (or I'm just ignorant).
Is language generation (following syntax and grammar and etc.) the hard part? Because personally, I'd be fine with a summary in the form of a series of small concept-token-graphs, as long as it was indeed a series of such graphs (akin to sentences, arranged into paragraphs), instead of one big jumbled graph.
Yep language generation is the hard part. What you have said is a stepping stone in that direction. Any task involving sequences is very hard today. How to generate these sequences ? Tops we can generate limited phrases, what Google smart reply does.
that is a nice idea, but i think we are unfortunately years away from that. i dont have time at the moment, but take this article and run it though this state of the art algorithm and see what comes out:
> the products have to come from "the approved network"[1]
This network also includes crowdfunding sites. Kickstarter's been promoting their involvement in this a lot for the past couple months. I'll probably see if this deal can get my recently-kickstarted graphic novel some extra promotion once it's shipping, since I'm already planning to use Amazon for fulfillment anyway.
Your thing has to be about a month away from shipping to consumers; if they allowed more of a lead time on that I'd have an application in progress.
1. Amazon Launchpad will ask for a few sample units to test. What they actually mean is that they sell these units. So, don't send Amazon units until you're ready to part with them and put them in your customer's hands.
Here's Vendor Express (their vendor management system) on that topic:
> We'll sometimes request a minimum sample of free units of your product so that we can evaluate the demand from our customers before placing an initial purchase order. After your product sells, we may start issuing purchase orders.
So, if you're a Kickstarter or Indie-go-go project and you want to sell your product on Amazon as well, you'll likely be going through Launchpad. This can get... uncomfortable for you if you send them test units before your Kickstarter backers get theirs.
2. Amazon's ToS for Vendor Express specifically states that they can rescind/change their Purchase Order at any moment before the product physically arrives at their fulfillment center. I've heard stories of startups having their P.O. reduced by half and having to eat that cost. When you're a young startup, capital is everything.
My take on this is that it sounds an awful lot like a sensing network for Amazon to discover which products are taking off and if it's viable to produce a low cost clone. At which point Amazon will not only have cloned your product, but also own your main distribution channel. Approach with caution?
However it may be a risk worth taking: on the one hand, you have a distribution channel that could help you take off, but might compete with you later. On the other, you'd have to run your own distribution, and it's less likely your product would ever take off in the first place.
> However it may be a risk worth taking: on the one hand, you have a distribution channel that could help you take off, but might compete with you later.
If you're worried about this, then you should also worry that competitors might compete with you anyway, whether or not you use their distribution channels.
The real question to consider is whether using their channel makes them more likely to compete with your or less, weighed against the other tradeoffs.
The best defense is to have a product that's hard to clone.
If the only reason competitors aren't cloning your product is that they don't realize there's a big market for it, you don't have a sustainable advantage.
No, you don't, but you do have an immediate advantage, that can be used to establish a userbase and obtain other sustainable advantages before other players get involved.
Amazon's treatment of their smaller partners (Android, Basics) has not been ideal, fwiw.
I too have wondered about this as well as how AWS might also be a signaling point for an analyst to point out other services Amazon might want to get into.
How different is this from store brands in the grocery store? He who owns the distribution calls the shots. The digital world was supposed to free suppliers from this with infinite shelf space but instead it just morphed into a slightly different and much more dangerous beast.
I think this is always true for any platform. With Apple you worry that your popular app might get so popular, Apple creates it themselves. In some ways, I think the point of platforms is to scale the vision beyond what you can do yourself.... until you can do it yourself.
The difference between the App Store and Amazon's model is that it's barely worth Apple cloning a competitors app because compared to their overall income, Apps will only ever be a rounding error.
Amazon Retail on the other hand, sell physical stuff. It's their bread and butter.
It's not evident what's new, but this initiative was announced a year ago, here's the article: "Amazon Takes On Product Hunt, Shopify With Launchpad, An All-In-One Marketing And Sales Portal" [1]
It's for just this sort of reason that I took my DSP plugin business 'Airwindows.com' to a Patreon-only model when my 'shopify' (Kagi) went bust recently.
It's getting pretty weird out there. This is very much about obliterating things like Shopify: helping startups is strictly tangential and/or temporary. It's all about the strategic moves and trying to prevent other players in the sector from growing. If they don't grow, they die, so seizing control is worth anything.
I've been using this for my dog toy that I launched on Kickstarter and it has worked out amazing. The only downside is you don't have control over what they price at since you're selling products wholesale to them. But for the amount of work that it takes (very little), it's a great trade off.
Are there any costs beyond (I assume) wanting to buy your thing at a wholesale price? I've got a graphic novel I just kickstarted and am curious about this program, but they won't take any applications that aren't about a month from shipping.
Amazon has got this weird thing with low-quality JPGs, and especially text-as-images for all their announcements. Every time they announce a new Kindle or something, the front page becomes a giant image of a "letter from Bezos" with artifacts the size of grapes.
It's constantly surprising to me how sophisticated and careful they are on every single other aspect of their business, except for this one.
Is there any chance that it could be intentional? Everyone knows Amazon's product page, for instance, is some of the most optimized real estate on the internet, I don't know any designer who looks at it and doesn't immediately want to start saying everything that is wrong with it. Could they, in this case, also be optimizing for something that we aren't considering?
Seriously. The testimonial block-quotes are images with meaningless alt text (e.g., "Testimonial 1"). They're clearly doing plenty of things right, but the inattention to detail and disregard for accessibility get my goat.
Because it's a jpeg and not a png.
Jpeg will show artifacts (smudging and mosquito noise) in high contrast areas and flat fields of color.
PNG is lossless, and well suited to compress repeating color values mixed with high frequency signals.
There are also wide quality variations across implementations. Kornel Lesiński has made some nice contributions to mozjpeg which improved quality significantly for images with sharp edges compared to the standard libjpeg:
Here is a similar version and size of the jpeg logo, that I created from a clean source.
It is about 40% smaller (in bits) than the jpeg.
http://i.imgur.com/DlJ2uFX.png
Again, I'm not disagreeing that PNG is better. I'm saying that, even if you insist on JPG, you can still do much better than what Amazon's got going on 90% of their splash pages.
I'm sure I'm not alone in having concerns about how ruthlessly centralizing Amazon is making the shopping experience for just about everything but groceries (because I will never trust anyone to pick my meats and produce for me), but if Amazon Launchpad works as well as this advertisement claims it will, then I cannot for the life of me think of a better means and marketplace to trial new products.
Maybe if your product does really well they'll set you up with a Dash button.
Although some of the points made by the blog are interesting, he is underestimating somethings. You don't need a buy button on the boxes at all, computer vision can auto-detect the brand and let Amazon offer you an option to buy it with little effort. You can do that today with barcode. Amazon already has all the infrastructure from "buying" to "shipping product to consumer". All they are doing is making buying easier - it's not revolutionary but a logical step forward. All said, retail stores should definitely be worried about this because Amazon is making it extremely easy to buy stuff now.
Especially the grocery stores, because I used to work in produce.
First thing they drilled into me was to bury the new stuff under the old stuff. Otherwise the old stuff gets and stays buried under the new, looks worse over time, and could even dissuade customers from buying anything on the rack at all if the old old stuff looks bad.
A store I go to usually has good produce but occasionally fails to take aging stuff off the shelf. Getting the picker to follow my subjective taste as far as what is too old is a bit of a problem.
Of course the solution is for the store to rotate the produce faster, but there is probably a reason they don't.
I know from a friend working on the project that this is a common sentiment about produce. I'm not sure if it's something they have addressed yet. For my preferences I think it could be doable:
How do you want your bananas? [a scale of 5 banana pictures from bright green to spotted... pick one]
You really think some kid working the summer shift, is going to meticulously look over every Apple going into your bag? Please. I will never do this, until it turns automated, with machines inspecting every single item.
This is an interesting idea in that it basically becomes the store for post-kickstarter projects that had traction but not enough to negotiate a contract with major retailers. Great idea -- I hope it takes off (no pun intended).
"Amazon Launchpad announced today the creation of a Kickstarter Collection (www.amazon.com/launchpad/kickstarter), featuring more than 300 Kickstarter products available for purchase on Amazon.com"
I used to sell hardware to amazon. Here's my take on this.
This boils down to a whitelist which allows companies to skip the traditional sales/retail-contract process required for a manufacturer to become a vendor.
This is not a service that Amazon will sell, essentially a loss leader to get vendors into the network.
Normally for brands that are new, Amazon requires a unique product line or price or some differentiator. A well established rep can easily get them in the door.
I typically recommend that brands:
1. Start selling online ASAP
2. Start selling on Amazon ASAP
3. As sales grow, become an Amazon vendor
Although, amazon is saying they welcome all startups but looks like this launchpad is only for companies having a physical product. I think you can't apply if you have an app to launch.
Your product has to be absolutely extraordinary to pass such an extensive list of vetting partners [1], in which case there's a good chance you would do fine selling on your own and generating your own word of mouth buzz if you ask me. If you invent the next iPhone you are probably going to do well even if don't sell on Amazon.
Doesn't seem that way to me, as a lot of the products that end up on Launchpad have a below 4-star rating, and they seem to currently have about 3,000 products listed. I also don't see how a lot of those products really deserve to be in the list.
I'm interested in the flip side of this. How, as a consumer, can I discover cool new products from startups that Amazon thinks are noteworthy? I wish the also announced an section of the website like amazon.com/launchpadproducts where I could see a list of qualified products.
They've been popping on the Amazon home page for me for about a year now, and usually one of the graphics will link to a larger list of launchpad products. I like to check it out every now and then see if there is anything interesting (it's how I discovered that Amazon started carrying Soylent products). This link seems to work to cut right to the listings: https://www.amazon.com/b/ref=br_imp?node=12034488011&pf_rd_i...
EDIT: Actually, there's a link right at the top of the OP's page titled "Shop the Store"
Interesting. I've just been reading about using alibaba with freight forwarders and shopify to create online shops where you can sell on Amazon. This might make it even easier.
It depends on what you're having manufactured. I was thinking of original products.
But I'd be doing it more to learn about supply chains, manufacturing, etc. and the problems people face. Plus, some of those sites support plugins and have their own ecosystems, so I might see opportunities for automation other people have missed.
One thing to note - they started out taking pre-orders but decided they didn't like this as they realized the startups have no control over their shipping schedule...
Imagine you're a startup/inventor that manufactures a widget that you'd like to sell to the mass market, you could either (1) try to sell direct from your website (2) get on the shelves of national stores like Best Buy, Target, Walmart, etc. The "shelves" include the real ones at their brick & mortar stores and the virtual shelves on their online websites.
(1) is hard to attract shoppers since nobody knows about your low-traffic website. Also, you'd have to handle the hassle of fulfillment yourself. Amazon Launchpad leverages their competency in global logistics to do this for you.
(2) is difficult to get meetings with corporate retail buyers and convince them to carry your product. Sometimes, there are also "slotting fees" (sometimes aka "bribes") to get prime shelf locations (eye level vs the floor.)
What Amazon is doing is opening up their "shelves" which includes the prime pixels real estate on their front page to promote startups' products. They are actively marketing your product. This is a different initiative from passively showing the 3rd-party marketplace sellers on amazon product pages.
However, to filter out the low quality junk and avoid every garage warrior trying to sell their flavor of homemade barbeque sauce, the products have to come from "the approved network"[1] that includes firms such as a16z, Accel Partners, etc. If we scan that list of affiliates, we'd expect the products vetted by them to be "cutting edge" and "innovative".
What's not specified in all the press releases and FAQ about Amazon Launchpad is the type of payment structure Amazon expects. Is it negiotiated on a product-by-product basis? Is it flat percentage?
[1]https://www.amazon.com/gp/launchpad/network
[2]https://en.wikipedia.org/wiki/Slotting_fee