Amazon and WalMart are not decision engines, and they never were. Maybe Amazon had a little of that on the side, but both are traditional retailers with all the traditionally hard problems the article mentioned that people don't want to solve.
In fact, Amazon is a poster-child example of what you stand to gain if you are willing to tackle more than just being a frontend.
Amazon started off selling books. What do you think their margin for each was?
Then they took over a large part of the market. Then they moved into selling other things like electronics.
Then they made the Kindle and controlled the device people read on and the starting point and the buying point.
Now they are getting into publishing and self-publishing.
They are taking over everything starting with taking over the online selling point for books.
That's what a fashion startup earning only 10% can do.
Use the 10% (which is mostly profit) to power an engine that slows takes over all of fashion.
Move into parallel areas like shoes and accessories.
Then starts making their own stuff.
I appreciate your trying to defend the poster of the original short-sighted article.
However, if you look at building a real business that has the power and control to completely take over an entire industry then starting online is much better than starting in the backend.
If you want to provide services to publishers who publish books (which is what back end would be for books) - well and good.
But there's nothing transformative about it.
The real transformation is to replace Publishers.
that's why the poster of the original article is so wrong.
She thinks she's thinking big by talking about tackling the back end instead of doing online sites.
However, the real prize is to replace the entire pipeline and then, if the company so chooses, to replace labels and the entire fashion industry.
Starting Point: Where people start their search (to buy something or find something).
Decision Engine: Where people make a decision.
Buying Point: Where people actually buy.
What are all the reviews at Amazon for?
Amazon is both a decision engine and a buying engine. They are trying to wean off starting engines like Google and that is why they first tried A9 and now are trying devices in customers' hands.
If you step out of your attempt to defend the poster of the first article you'll realize that her article (not her, just the article) is very short-sighted.
The Internet is destroying entire industries - yet she thinks fashion internet startups should instead compete in the real world.
That makes no sense. We should all play to our strenghts. While she's trying to do backend stuff and help existing Fashion companies, some 'only 3 to 8%' earning Internet startup is going to disrupt the entire industry.
A hacker, in my opinion, is not meant to figure out how to solve the problems the existing industry power players created or want solved.
A hacker, again in my opinion, should solve the most elegant things and solve things for actual people.
You are missing a very key point here: consumers do not decide what they wear, the fashion / PR machine decides that for them.
Think about this: while a girl may discover a cool, new designer on Pintrest and buy a dress from that designer, trust me, she will still scratch out some girl's eyeballs to get the new Louis Vuttion bag. And why? Branding. And hundreds of millions of dollars in marketing to create an image, an illusion, and an object that conveys status. LVMH is the most powerful and profitable luxury conglomerate in the world for a reason (it is about 3x more profitable than Amazon with 20% less revenue).
Brands either die or thrive based on their ability to do two things: create a desirable brand and manage inventory. The entire function of branding and marketing is not going to be replaced by Svpply. Or Pintrest. Or whoever.
Lastly, it's comical you keep mentioning Amazon. One of the biggest reasons for their success is their ability to manage a supply chain better than almost anyone else.
I've found that almost nobody in Silicon Valley gets this, especially engineers (speaking as one). They treat the "product problem" as a technology problem (more data, better recommendations), but it's really a psychology problem (how do you make people fall in love?).
Mapping the social web onto the latter problem is actually more straightforward, IMO, than trying to, say, mine my Facebook account to make more "personalized" apparel recommendations.
In fact, Amazon is a poster-child example of what you stand to gain if you are willing to tackle more than just being a frontend.