Let's assume that you would need 1000 machines to have as much throughput as 4000 people.
WAG, but I'd wager that each of those machines would cost $500k.
That means 100 million in immediate outlays. WAG 2: fuel/electricity/maintenance is $200/month, and you need a team of 20 engineers to watch over them at $10k/engineer/month. So recurring costs come to $400k.
Which brings us to: 100M/1.3M = 77 months, or between 6 and 7 years.
That, however, doesn't take into account the opportunity cost of the initial $100,000,000 outlay. That brings it up to around a decade.
Labor costs are likely double, if not more, than what i've estimated. overtime + insurance + drug testing + security + all the other crap i'm forgetting.
Well if you treat people as human machines, remove unionizing ability and what have you, then you can bring yourself to competitive levels with China, which has kinda made a very persuasive point that human robots are cheaper than mechanical ones.
The only issue with labor of course is the so called 'managing' aspect of it, which covers things like quality of life. Robots don't have and will never complain about, while being able to do tasks at a level that most humans wont ever be able to.
Re: " China, which has kinda made the point that human robots are cheaper than mechanical ones."
Not for long.
"The China Business News on Monday quoted Foxconn Chairman Terry Gou as saying the company planned to use 1 million robots within three years, up from about 10,000 robots in use now and an expected 300,000 next year."
The point still stands. It may (should) change in the future as people start asserting their rights. (I have seen the article before, its Foxconn currently, but labor prices and further opportunities haven't reached a point where you can safely bleed off the population away from manufacturing just yet.)
At the same time though, it shows that as long as you have people who have no option, you can use them to easily produce more value than robots at similar costs. Which is what the Mother Jones article is basically about.
Also, its worth remembering that those robots are also going to be used for capacity expansion, while keeping cheap tractable labour.
TLDR: Improving standards of living will make robotics more competitive, unless there are sufficient people who have no other option but to compete with machines.
Most companies don't have 4000 employees, or need 1000 robots. Like every business, you add capacity as necessary. Diapers.com started with one mom and minivan.
This article: http://money.cnn.com/2011/11/09/smallbusiness/kiva_robots/in... says that 1000 robots costs $15-20 million, not the $500 million that 500k a piece would run. Also what is there to indicate such a hefty need for engineering oversight? 20 engineers constantly overseeing the machines sounds like they are either constantly breaking down, or need a significant amount of custom programming per unit as upkeep, both of which just sounds like a quality control issue on Kiva's side. It makes sense that there's a lot of initial planning/programming for each installation, but I'd expect it to be mostly self-running after that. I would be curious about the energy needs per unit, though. They seem to be carrying around more weight per package shipped than with a human picker carrying just what goes in one package, but then again their use of energy might be more efficient than humans.
1000 robots needing just 30 minutes of preventative maintenance per week is 500 hours of PM a week. That sounds like about the amount of PM a staff of 20 engineers can supply, once you account for admin, PTO, training, travel between robots, and doing the actual work.
Engineer does not solely mean "one who creates software".
How much automation do you need and how "bulletproof" does it have to be? What's the up time? What sort of rates does it have to handle average and peak? How accurate does it have to be? Does your control and execution software have to be custom (since you business model and methods are oh so different than your competitors) or will the standard be good enough? Trust me automation and the software to drive it can be quite expensive.
I'm really curious about the robot costs. The automated shelving systems i've seen look, well, really simple. I'm pretty sure i could get a half assed DIY system working for a few hundred dollars. An industrial version can't be more than 100x the price can they? 50k per mover, on the outside?
> "I'm pretty sure i could get a half assed DIY system working for a few hundred dollars."
you think the same thing when you see an Oracle instance that cost tens of millions and think you could have done it with postgres. what these retailers are really paying for is reliability, redundancy and support - having a big brand co. to call when something goes wrong.
that isn't to say that there isn't a mysql/pgsql style opportunity within warehouse automation and supply chain management - it is just that it isn't likely that Wal-Mart and Amazon would be your customer. Similar development and deployment cycle as with mainframes and servers - what was once the territory of only large companies and governments is now an accessible technology and competitive advantage amongst small and medium businesses as well
The most interesting aspect of supply chain management to me is the concept of a completely outsourced warehouse - where it is cheaper to have a specialized company manage and run your inventory and supply chain as part of their larger infrastructure (and economies of scale etc.) rather than building your own warehouses and system. Amazon became very very good in this field because an outsource style solution didn't exist at the time and they had no choice other than to do it themselves, but you could imagine that an Amazon being started today would not have its own warehouses and would not be writing long letters to shareholders trying to justify hundreds of millions in capital expenses in order to automate warehouses and bring down margins.
The three examples often cited are Amazon, Wal-Mart and Diapers. Amazon implemented a lot of their own systems using partners, Wal-Mart has invested billions and Diapers.com implemented with Kiva Systems (amazon now own diapers).
Amazon and Diapers could automate end-to-end because they had fixed product sizes and packaging. A lot of other retailers like Wal-Mart are attempting to paletize their goods for this reason. It is very expensive to completely automate end-to-end with retailers who have a broad inventory (which is why the vertical online retailers such as Diapers and Zappos did so well, they could lower margins with an easier to manage supply chain).
I used to follow Amazon stock and filings. They spent hundreds of millions of dollars on automation in their initial warehouses. They spent so much on the servers that run software controllers that the RedHat stock got a big bump when it was announced that they were the partner implementing it (Amazon made up a double-digit percentage of RedHat revenue).
I remember headlines of Amazon investing ~$100M into updating single warehouses. A lot of time was spent analyzing the outlay and returns - it was definitely a long-term investment rather than something you can immediately identify as being more cost-effective in the short term.
So if you have varying inventory and demand cycles it is less cost-effective to automate supply chain. There is also the part where you need to integrate with backend systems - SAP, Oracle, Sage, etc. which again involves consulting time and multi-million dollar projects.
From this story, it sounds like this warehouse as a very broad inventory. For eg. one bin contains batteries mixed with DVD's etc. which isn't suitable for robotic system since all they do is grab the basket, knowing what is inside it, and bring it to the packaging conveyor. It sounds like they already do this for the most popular products, and it is the rarer products where having a dedicated area for its stock just isn't feasible. This is also why Wal-Mart went straight into investing billions into RFID rather than the barcode scanning model used by Kiva.
IIRC, just the automation robot hardware market alone is ~$3B p.a, and online commerce with goods is ~$30B p.a, so already 10% of revenue is being invested back into hardware alone, which gives you an idea of costs and limitations. It may be a market that is ripe for disruption, since the deployment model seems to be similar to how large backend enterprise systems are implemented with Oracle, IBM etc. there doesn't seem to be any solution at the low to medium end of the market, although that is part of Kiva's pitch as well (they have standardized robot and bucket sizes).
The whole area is really interesting, I have tons of bookmarks on another laptop if you want me to send them to you. I looked into it some years ago as part of just analyzing tech companies and their margins (my main takeaway was narrower inventory, vertical market = better margins and better automation, and that it gets very expensive for broader inventories). To find more, a good starting point is the companies that sell the hardware and implement the systems such as Kiva: http://en.wikipedia.org/wiki/Kiva_Systems. They must have a good PR department because their customers and implementations have been written about a lot in Wired, the WSJ, NYTimes etc. as the future of warehouse automation, for eg.
Amazon probably knows their business better than you do. If they're hiring people instead of buying robots, that's probably because people get the job done cheaper - for now.
WAG, but I'd wager that each of those machines would cost $500k.
That means 100 million in immediate outlays. WAG 2: fuel/electricity/maintenance is $200/month, and you need a team of 20 engineers to watch over them at $10k/engineer/month. So recurring costs come to $400k.
Which brings us to: 100M/1.3M = 77 months, or between 6 and 7 years.
That, however, doesn't take into account the opportunity cost of the initial $100,000,000 outlay. That brings it up to around a decade.
So, a decade to break even.