Take, for example, a traditional retailer such as Opening Ceremony. After shipping, merchant fees, packaging, and COGS, they produce an average gross margin around 40% (the industry standard) – after accounting for photography expenses (which most affiliate sites do not have), let’s say the margin is around 30%. Whereas a fashion site that generates revenue mainly from affiliate fees will collect only 3-8% of the revenue on each sale. That means that the new, lightweight fashion site must sell 4x-10x more inventory than Opening Ceremony in order to produce similar profit margins. On the hierarchy of risk, figuring out a way to sell 10x more than your competitor in order to just stay in the game is a much larger risk than the inventory management issues of a traditional retailer.We're seriously comparing B+M company with gross margins of 30% to a lean startup receiving affiliate fees of (let's take the worst case, even) 3% on a product and suggesting that the affiliate has to sell 10x as much "in order to stay in the game"?
Plain and simple, you cannot compare these numbers, they have absolutely nothing to do with one another.
That 3% affiliate fee is as close to pure, unadulterated gravy as you can get in business - it's profit practically from day one, regardless of volume. It's probably being chopped up between three dudes working out of a garage somewhere, whose entire set of business expenses comes down to some electricity, a few meals a day, and a $0.34 / hour large EC2 instance. If their volume went up by a factor of 10, then maybe they'd need to add another few servers (each of which is probably making them thousands of dollars per hour, if it's pegged).
With affiliate marketing, if you've got high volume, your margin approaches 100%, since your expenses are pretty much independent of your sales (unless you're reliant on advertising, which to be fair is another matter altogether) - would it then make sense to say that traditional retailers need to sell 3x as much product to stay above water? No, because it's an apples to oranges comparison.
Now, if you wanted to claim that in order to have similar total profits to the traditional retailer the startup would have to move 10x as much product, we can start talking. But the point is, they don't need as much profit to compensate everyone involved at the same level - that traditional retailer probably has at least 10x as many people employed per unit sold as the startup does, and the startup's advantage there scales much better with increasing volume.
It certainly may be the case that pure-online fashion sites are fundamentally doomed to fail for some reason, but if so, the article offers no real evidence. All we've seen is an argument that would also imply that Amazon, Netflix, iTunes, and most other online product sales businesses should fail, because they all have much smaller profit-per-unit figures than traditional brick and mortars.