My Reply In Red
Let me start by saying thankyou for writing this article. My aim of writing this is not to get in to an argument with you about why drones are good for deliveries or not. I just think I have some good answers to the questions that you raise. I have a prototype of a delivery drone that address a few of the problem that you see with drone delivery. You can find more information on this site.
Disconnect: drone delivery is not the future of ecommerce
Drones. What started as a science fiction trope has become a cool tool for film hobbyists and a terrifying weapon for the world’s militaries. But the real potential for drones is in ecommerce, or so many people in the tech industry have been saying. And ecommerce companies are listening, too. In the West everybody from Amazon to Wal-Mart is looking into drone delivery, and here in Asia Alibaba and SingPost have already conducted drone delivery trials. It all looks very cool.
But the simple fact is that wide-scale drone delivery isn’t happening any time soon. And by “any time soon”, I mean “ever.” Sure, ecommerce firms will probably start using drones more to get lighter-weight packages to remote locations difficult to reach by car. And they might use drones internally for tasks like sorting and moving products around a warehouse. But this idea that Amazon or Alibaba will be dropping packages at your doorstep with drones? It’s never going to happen. Here’s why.
Infrastructure
Drones have two major advantages over the traditional van-and-driver system of package delivery. The first is that they’re autonomous. The second is that they can reach places roads don’t go. But with autonomous vehicles probably closer to hitting the roads than delivery drones are to really hitting the skies, automation isn’t an advantage for drones – cars will have that too.
Cars do require roads, of course – drones have them there – but the fact is that in most countries, there are roads going to most places. You probably live on a road. It’s probably connected to a network of other roads, bridges, and tunnels, all designed and administered to move goods from one location to another efficiently. There are police on the roads to ensure safety for all, and laws and court systems trained to handle the fallout when things do go wrong. Autonomous delivery cars and trucks can take advantage of all of this easily. The infrastructure is already there; it’s really just a question of getting the software right (and maybe giving the cops some sort of kill switch so they can shut down runaway autonomous vehicles remotely if need be).
In contrast, there’s no infrastructure at all for drone flights. That is the beauty of drones, minimal infrastructure. There are no systems in existence to regulate where drones can fly, current air traffic control is an example of a system in existence that can regulate aircraft. DJI already have software that prevents drones from flying over know airfields. This is a very simple software upgrade to a drone that DJI implemented after a drone crashed in to the white house lawn. for example, or deal with the fallout from accidents when they do occur. There is no police force in place that could oversee and regulate drone delivery. Drone deliveries will be done by the big boys, google amazon etc. It will be in their best interest to comply with any and all rules. Any rogue delivery companies will be found by existing authorities and penalised. All of these systems would need to be built. Legislation and regulation would need to be written. Yes, I agree, and these things are being hashed out as I write this. One of the recommendations is to have drone registration. And yes, we could build all this new infrastructure, but we almost certainly won’t bother because…
Expense
Getting stuff off the ground is pretty expensive – this is something you’ve probably noticed if you’ve ever bought a plane ticket. Usually (not all the time, but usually) it’s cheaper to drive or take a train. Whether this will always be true depends on a lot we can’t be certain of, like what the source (and thus cost) of energy will be in the future. But we know that broad-scale ecommerce drone delivery would mean delivering much heavier packages than what’s been used in ecommerce drone tests, at much long ranges than what’s been tested. Jeff Bezos said in his interview about drone deliveries that, their drone will be able to carry 2 pounds and 80% of the items sold on Amazon are under 2 pound. Meaning that up to 80% of deliveries could be done by drone. Heavier items would use the traditional way. SingPost used a letter and a T-shirt, Alibaba used packets of tea, and both companies ran their tests over very limited distances. But if drone delivery is going to become the norm, then these things need to be able to deliver flatscreen TVs and bed frames and dumbbell sets, I will admit that I can’t see drones delivering anything much over 5kg and they need to be able to fly for miles and miles. That means bigger, badder drones. What those sorts of drones would cost in an ecommerce setting is difficult to estimate, but the military has drones kinda like that that cost US$2,500 and up per hour of operation. It cost between 20 – 70 cents to transport a 2kg package 10km. A 5 fold saving over standard truck delivery (source Matternet.com)
I don’t think Amazon Prime is going to cover that.
But even if we ignore that issue entirely and assume that moving one ton of product via drone requires less energy than moving that same ton of product the same distance via a wheeled vehicle, there’s still the question of paying for all of those drones. Drones are becoming surprisingly cheap. The kitty hawk heavy lift drone can lift around 4kg. It costs around $5000. Compared to a delivery truck ($25000) that is extremely cheap. I understand that a truck can lift much more but a drone can travel much faster for less so a drone could possibly deliver just as much over the same period of time, especially if that truck had to deal with traffic.
Cars, as we know, already exist. There are an estimated 1.2 billion motor vehicles on the world’s roads already, and that number is expected to climb. The factories required to produce these vehicles in such great quantities are already in place, and logistics companies already own the motor vehicle fleets they need for deliveries. Implementing automated vehicular delivery is a cost saver – you fire the human driver, outfit the existing trucks with a little new hardware, plug in your software, and you’re good to go. I think this view is a little too simplistic. Driver less cars are coming and I am really looking forward to it but simply plugging something in to your truck and away you go is a way off. I expect drone deliveries to be wide spread before retrofitting existing truck fleets will be that simple. With drones, you get the same driver savings (any way you slice it, the future looks bleak for human drivers), but you’ve got to buy an entirely new fleet of vehicles, you’ve got to train or hire a new set of mechanics to fix them when they fall apart, fixing drones is very much a plug and play thing and that is only going to get more simple in the future you’ve got to set up new recovery procedures for when something goes wrong, etc. It’s a much costlier up-front proposition. Recovery of standard delivery trucks will be the same as recovery of drones. The advantage is that the drone will only be carrying 1 package. The truck will be carrying many packages meaning that many deliveries will be delayed not just 1 delivery.
Maybe technology will magically solve all of that, somehow. But even if it does, there’s still the issue of
Cities
Right now, if you live in a city when you order something from Amazon or Tmall or wherever, what happens? A guy in a jumpsuit shows up at your apartment door, inside your building. He probably had to navigate doors and stairs or operate an elevator to get there. He might even have had to sign in or interact a doorman if you live in a high-end apartment. Drones can’t do any of those things. With our system they can. www.blktatu.com It employs a delivery catchment system specifically designed for cities.
That means that for people living in urban environments, drone delivery would actually be less convenient. Instead of getting packages delivered straight to your door, you’d likely have to go to a local “drone pickup location” and lug your stuff home yourself. That is, if it ever got there, because there’s also the problem of….
Theft/Damage
Drones are expensive. Delivery drones will be carrying whatever brand new stuff you ordered, and likely flying over other people’s property to get to you. That’s going to make them a pretty tempting and easy target for thieves. Even if you put trackers in them, there’s no way you could respond to every errant drone in real-time. Currently in many developed countries packages are left on door steps with traditional delivery methods. Theft is not considered a big problem and insurance is employed if theft is a concern.
Thieves aren’t the only problem, either. Kids would be throwing rocks at them just for fun. Hunters would be taking potshots at them for practice. Hawks would be attacking them. Harsh winds, rain, and hail would be blowing them off course or destroying them completely. You might even have to deal with black hats or rival companies hacking them or shooting them down to gain a competitive advantage. There will be times when people have a potshots and kids will be throwing stones at them but the majority of the time when the drone is flying it will be high enough to make it an extremely hard target. There is nothing stopping people from sabotaging autonomous vehicles and raiding the contents of the van. It would actually be much more worth their time to attack a van because there are more packages on board and more opportunity to find something valuable to them. There will also be no fly days where the weather will not permit flying. These days will also exist for ground transport when there is flooding or heavy snow.
Trucks, by comparison, are pretty safe. They’re rock- and hawk-proof and rain-proof, and they’re a lot harder to shoot at or steal without getting caught.
Complication
At this point it’s probably obvious that drones aren’t ever going to be the one-and-only delivery logistics solution for ecommerce firms. But could they employ drones for some deliveries, which sticking with traditional logistics solutions for others? I agree that they will not be the only solution.
Maybe. But managing two logistics systems (traditional and drone) at once is more complicated, and therefore likely more expensive, than just managing one logistics system. If you’ve got two systems then you’ve got different types of vehicles that need different kinds of maintenance, likely different energy sources, different package preparation standards, different accident resolution systems, etc. Setting up all of that stuff once costs money. Setting it all up twice costs more money, probably, and it also costs time and mindspace. Logistics already employ many different forms of transport to get stuff around the globe Air Sea and Rail are just a couple. Adding another one that reduces costs would not be a hard decision to make. That doesn’t necessarily mean it won’t be worth it, but it’ll have to be very worth it for companies to implement two logistics systems on a broad scale. Nobody likes introducing complications just for their own sake.
