These were clothing style pickers, and it sounds like the CEO intended to get rid of many or most of them.
It wasn't just quitting. They were offered $1000 in severance. Only a third of the workers took the offer.
It seems like they're trying to get consistency from fewer staff that are more focused on the job. And less costly to employ. That, or eventually replace them with algorithmic picking.
The company missed revenue targets, lost multiple senior staff, and the new CEO is from Bain capital. They're doing this on purpose.
That explains the introduction of black-box management-solely-via-P&L. If any position could be automated away, one would think that that sort of CEO could...
Maybe that's how we solve capitalism, we use machine learning to replace the capitalists!
I can't help but imagine that Neutral Planet from Futurama would be the end result of that, constant model fit of the median until the market doesn't even cater to the outliers, and everyone wears the same grey jumpsuit.
Quite right, one might as well speak of replacing vampires with werewolves: it still won't be safe to walk at night! Of course the point is to be rid of them all.
I think the CEO is missing a major point. Their competitive advantage might be that they AREN'T algorithmic.
Shopping algorithms seem really good at finding very similar items in my experience. That's great when I'm shopping for a cheap router, or some other commodity. When shopping for something like clothing, you frequently want VERY dissimilar items to compare, OR you want complementary, but differently categorized items.
Get rid of the human touch at your own peril. Especially when your business model is built on taste.
Looking at stitch fix pricing it is not cheap. It looks a lot like you would expect somewhere like Nordstrom's to price. It's worth noting that Nordstrom's is more than happy to have an employee act as a personal stylist at no cost.
I have a dark suspicion that their best target market might be to pretend they're not algorithmic, but actually being so.
Most people, almost by definition, have rather ordinary tastes. They will want to imagine that it's slightly better than average, but not outrageous.
I don't know much about fashion, but I see it a lot in food marketing. Restaurants differentiate themselves on trendy but safe choices. It works down from the high end: Gordon Ramsey splashes something with truffle oil, and a few years later you can get New Burger King Truffle Fries. A lot of "mom & pop" restaurants are just heating up things off of Sysco trucks.
I don't mean that to sound snobbish. People should enjoy whatever they like. I'm slightly turned off when it's marketed as being really innovative while smaller, more interesting things languish, but that's just the market at work.
As it applies here, I suspect that most people really could be very happy with algorithmically applied clothing. A truly personal stylist would be much more expensive and outside of most people's comfort zones anyway -- unless they did basically the same thing as the algorithm.
What Stitch Fix can offer is the illusion that you're being truly stylish without any of the risks. Which is a fine thing, as far as I'm concerned, if it makes them happy. That excludes a market which really does want a truly personal, human stylist, but I suspect that market isn't nearly as large, and even smaller if you ask them to pay what hours of attention would really cost.
Adding a very thin veneer of humanity on top would be a big win. Have the AI do most of the work, then have a human being sanity check it and add a personal note (maybe something derived from their conversations... which could itself be mined by the AI and then massaged by a human).
If it's nothing more than "You mentioned that you liked the coat that [celebrity] wore, and this was a similar sort of look but in [color] you said you liked", that could potentially give people a really positive experience. Even if many people are wearing the same coat because the algorithm recommends it.
What sort of costs go into these employees? I think I'm missing something, but not very familiar with costs of a worker who:
* has no benefits
* BYOD
* not eligible for benefits (they are explicitly called out in TFA as not eligible to become employees)
In the long term, they cost more than the code that is just optimizing delivery of what's in the warehouse to "what each consumer has been surveyed as preferring/purchasing".
This seems like an inflection point for the company where their software is starting to be more useful than their employees. It's the equivalent of McDonald's replacing counter people with kiosks.
It wasn't just quitting. They were offered $1000 in severance. Only a third of the workers took the offer.
It seems like they're trying to get consistency from fewer staff that are more focused on the job. And less costly to employ. That, or eventually replace them with algorithmic picking.
The company missed revenue targets, lost multiple senior staff, and the new CEO is from Bain capital. They're doing this on purpose.