E-Commerce Shipping

Are Robots the Future of Last-Mile Delivery?

Fri 09 Jun 2017
By Shippo

Milkmen, newspaper boys, bicycle couriers, drones, robots - humans have been looking for a way to deliver goods to customer’s doors for centuries. As technology evolves, so do the available delivery options.  One of the most interesting new delivery technologies is not on the roads or in the air. It is semi-autonomous delivery robots on sidewalks, a quickly growing industry.

Hear about how these robots can solve real social and economic problems while making last mile delivery more efficient, cost effective and convenient for customers and businesses alike. We invite you to watch this presentation by Henry Harris-Burland from Starship Robots from the Shippo Summit. If you prefer, we’ve also transcribed his presentation below.

I'm here to talk about delivery robots. Not just our delivery robot, but actually the wider industry. What we saw over the last couple of years, what the problems are, and what to expect over the next couple of years and what this technology can bring to logistics, but more specifically last-mile logistics.

We are trying to solve the last-mile delivery problem, the last couple of miles between a hub or a store and a customer's house or place of business. It's notoriously inefficient at the moment. Up to 40% of the total cost of the transportation can be the last couple of miles. Which means if we're ordering something from China, the first 4,900 miles might account for 60% and then the last two or three miles account for the remaining 40% of the cost. It's super inefficient. Businesses are trying to make their last-mile-delivery logistics much more efficient and cost effective, but it's very difficult. Those are the problems that we're trying to solve.

Last Mile Delivery Challenges

As an example, vans stopping outside 100s houses every single day, that's traditional last mile. But one thing that's often regularly overlooked in last-mile delivery is personal shopping trips. That's you and me getting in our car. I'm guilty of this very regularly ... to go a mile and a half down the road to pick up two or three bags of groceries, and then move our one-and-a-half-ton car back to our house a mile and a half down the road. I do this all the time. It's a part of last-mile logistics. It's a part of delivery that actually hasn't really been monetized yet.

Delivery robots can greatly assist with saving us our time. It doesn't really make sense if you think about it. Why are we wasting our time getting in a car to travel just a mile and a half down the road to pick up some groceries when technology can now do that for us? We're looking to solve those problems as well.

We're all customers of parcel delivery, groceries, takeaway food. We always want it free. We want to be able to return things easily, same day. It comes down to cost. It comes down to time. It comes down to control.

Now, it's very difficult for businesses in traditional logistics to be able to offer this. Many companies are seeing a huge gap between customer perception, and business supply. There's a massive gap at the moment between what we want and what they can offer. Anybody that's used Amazon Prime and Amazon Prime Now knows it's such an incredible service that. If you've used it once or twice a week or once or twice a month, you can get used to that level of control and speed very, very quickly, and you want it when you're ordering anything. Businesses can't offer that yet, and it's because it's not profitable. For a lot of businesses, it's not sustainable.

Delivery robots, not this second but in the future, are going to be able to fill that gap, to offer what customers perceive that they deserve right now, because we always want more. The problem is the last-mile math. It's super simple. Labor, vehicle, overheads, divide by number of deliveries, equals your cost for delivery. Now, there's two major costs there, labor and the vehicle. It's very, very difficult to optimize those costs down. We're looking to dramatically reduce the vehicle cost, dramatically reduce the labor cost by introducing delivery robots. That's how you can get to very, very low prices for on-demand delivery.

The $1 Delivery Goal

We are ultimately aiming for $1 on-demand delivery in the future, and that's totally unheard of at the moment and actually not very possible. What businesses are doing at the moment is there's one business ... I won't mention any names ... that only takes right turns at traffic lights, so they don't have to wait at traffic lights. They only take right turns. They're squeezing every single percentage of efficiency they can out of their current business model, and it's great. They might be getting 1% or 2% efficiencies, single digit, 4% or 5%, but the main problem still stands.

E-commerce is growing at 9% or 10% every single year. That quite simply means you and me are ordering more and more things online. That's more deliveries. That's more vans. That's higher pressure and a strain on resources. Something has to change, whether it's now, whether it's five years, whether it's 10 years. We are offering this technology now, and it's a very exciting time because we can see double digit efficiency increases that previously were not possible. Here's our solution. It's delivery robots. It's replacing humans and vehicles that may be doing those last one or two miles with semi-autonomous delivery robots.

Now, these robots aren't a gimmick. Some people see these and think, "Oh, that's a bit of fun. It will never work." It's a fair assumption on first impressions, but we are actually trying to solve real social and economic problems here. We go to cities. We talk to cities, and they understand that they may have big problems now and in the future with congestion and pollution. We're looking to take those cars, take those vans that are stopping and starting hundreds of times off the road and replacing them with zero-emissions delivery robots.

We're looking to take those cars, take those vans that are stopping and starting hundreds of times off the road and replacing them with zero-emissions delivery robots.

Now, some people would then say, "Well, you're going to take the congestion off the road, and then you're going to bring it onto the sidewalks." These robots travel at four miles an hour on sidewalks. They act like pedestrians. They cross the road like pedestrians. They don't go on the roads. They only go where pedestrians go, but you're not going to suddenly see a huge row of delivery robots on sidewalks. They're not going to become a big inconvenience in our daily lives. Otherwise, they wouldn't be successful. Even at peak times, delivery robots would become a common sight, but by no means are they going to become an inconvenience, and that's very, very important.

There's other areas that delivery robots can assist cities and also businesses, the elderly, the less abled, the disabled that might be reliant on public transport. They may be reliant on family members to get to the grocery store or go get their food. Delivery robots can give those people their power back by offering on-demand, 15-to-30-minute delivery to their door for $1, maybe a little bit more, maybe a little bit less, but our ultimate vision is $1 on-demand delivery.

Another area where delivery robots can greatly assist businesses is at a local level. Right now, at Starship, are working with relatively big businesses. We're working with DoorDash, Postmates, Domino's, Hermes Logistics, Swiss Post, to name a few, but in the future, when we get that cost down to $1 ... which we're not at yet, I admit, but that's the aim. When we do get that cost down to $1, we can open up this service and other delivery robot companies can open up this service to local businesses that can't offer delivery right now.

If you're a local bakery, a local coffee shop, it makes absolutely no sense to offer your coffee and croissant delivery a mile down the road. You'll go out of business within 24 hours because of the cost of delivery. If you can have access to a service that offers delivery for $1 on demand, you can get on a level playing field with the big multi-million dollar businesses. We've found tremendous pickup. Even though we can't offer that service yet, we've found tremendous interest from local businesses in all of the jurisdictions that we're offering.

These are some of the technologies that I'm sure a lot of us are familiar with. Again, lots of people think this is five, ten years away. I think it's worth making a distinction, a very clear distinction. Commercially available technologies like you see here might be quite far away, might be five years, but these are testing and doing real deliveries or real trips with customers right now. There is a big difference between a pilot program, which is what we're operating, and a rolled out commercial service. I think it's worth differentiating the two. All of these technologies you see are testing and piloting right now, but none of them are commercial services, and we may not see that for a couple of years at least.

Semi Autonomous

One big thing I want to mention here, one big difference is the percentage of autonomous driving. I'll come to it later on as well, but it's very, very important. This robot and other delivery robots are only aiming for 99% autonomous driving. Now, there's many reasons for that, but we want an element of human oversight and human control. There is always going to be a time that a robot finds itself in a particular situation that it can't do on its own, and it would ping up to a human operator, who could very quickly take control, make a decision, find an alternative route, or whatever the case may be.

It's really important to note we never want a 100% autonomous device, and that's another reason why you're likely to see this technology before others. If you're a self-driving car, you need 100% autonomous driving before you can roll it out commercially. You don't have the luxury of driving at 80 miles an hour down a freeway and ping up to a human operator to quickly decide if that's a bag full of bricks or a bag full of air pinged up in from of you, and you don't know whether to swerve or go straight over it. You don't have that luxury at 80 miles an hour, but if you're trundling down the sidewalk at four miles an hour, you do have that luxury. You can come to a safe stop very quickly. You can ping up to a human.

Now, of course, the robot has very sophisticated obstacle detection so, in reality, situations like that, it would be able to solve itself, but there are many, many other situations that it may not be able to solve itself, and that's where this level of human oversight is really, really important for this technology and why we can, and others, do deliveries right now to real customers, real deliveries to real customers even though it's under the umbrella of a pilot program.

What about drones?

Again, aerial drones, other technologies. Lots of people ask me, "Aerial drones must be a competing technology to ground delivery robots." Again, it's a very fair assumption. Our opinion is that aerial drones are very well suited to particular use cases. Aerial drones are fantastic for rural deliveries, deliveries over deserts, deliveries to farms, deliveries over rivers, forests, delivering medicine to Rwanda like some of the companies actually are doing right now. That's fantastic for aerial drones. Humans find it difficult to go to those places. Robots would find it very difficult to those places, just like cars. Aerial drones are perfect for those sorts of situations, but if you come into more suburban and urban environments where people live in lines of residential houses and flats, that's where delivery robots on the ground are much better suited. They're more efficient. They're more cost effective, and they're safer.

At the end of the day, it seems as if people aren't necessarily happy with aerial drones buzzing above their yards when their kids are out playing in the street and things like that. If something catastrophically went wrong, it will fall out of the sky. If something catastrophically goes wrong with our robot, everything suddenly failed, it would come to a safe stop. There are differences. It takes about 10 times less energy to roll something along the ground than it does to lift it up. Finishing that point, complementary technologies, I would say, rather than competing.

The Delivery Robot Industry

There are around about 10 delivery robot companies now. We were the first, but there are some in China, some in Silicon Valley. There's Alibaba getting involved. Big companies are moving themselves into this space. It's a really interesting time for this industry. We're at the very beginning of the industry right now, and I compare it to the self-driving car industry say four or five years ago, when people were just beginning to hear about it, and now, depending on what news you read, we hear about it every single day there's a news story. We think it's going to be the same for delivery robots.

Hybrid Approach

There are other technologies in the space at the moment. Some of you may have seen the Robovan. This is actually a partnership that we have with Daimler, Mercedes-Benz vans, incredible piece of technology. Put it quite simply, it is a parcel delivery van filled with robots, and these delivery robots are loaded in the van. There are crates where the parcels are above the robots. It uses a pick-by-light system, so the driver knows what parcels to put in each crate and which crate to put in each robot. This would roam around the streets of a particular neighborhood, dropping off robots to deliver, and also picking up empty robots to reload.

Robovan - Starship and Mercedes Benz Partenership

We've seen, through our calculations, huge efficiency gains when using this type of technology, because vans are still the most efficient method of delivery over the first 10 miles, 15 miles, whatever it might be. Robots are the most efficient method of delivery over the last couple of miles. Robots have a particular area where they are most efficient. You're not going to get a pizza delivery with a robot if you live 10 miles away from the restaurant. It's going to be pretty cold. These robots deliver within one to two miles. We aim for delivery in 15 to 30 minutes on demand. That is the ultimate aim.

Industry Focus and Operational Plans

We focus on three industries: parcel delivery, groceries, and takeaway food delivery. We aim for delivery in 15 to 30 minutes, as I said, and you would get a robot by ordering food through a partner that we work with. If you were to download the DoorDash app or the Postmates app right now and lived anywhere near Redwood City or San Carlos or Sunnyvale, you could receive a delivery with a robot if you lived within a two-mile radius of the restaurant.

This is a platform as a service. We're not selling robots. If you were a business, you'd be paying per delivery, per mile, or if you're a big business and you wanted thousands of robots, you'd be paying a monthly fee, almost like a lease. If there was a fleet of robots within your neighborhood, it would be a service that you could just very easily plug into, and your robots would be turning up to your door. We try and make it as simple as possible for new businesses to be onboarded.

One other thing worth noting and another reason why we're not selling robots is they would be a shared fleet. That's another reason why you don't see branding on this robot. If we were to work with a business, we're not going to suddenly put stickers all over the robot. No one's ordering a pizza at 7 a.m. ... or I speak for myself. I don't know of many people that order pizzas at 7 a.m. I don't want to judge. People are getting parcel deliveries though, maybe before work. They may be getting grocery deliveries. Lots of people are ordering food at midday and at dinner. Again, these different industries that we operate in have slightly different demand peaks, and it's interesting to note that these robots, in order to maximize utilization and maximize efficiency and therefore decrease cost, they should be shared fleets being used throughout the day. That means you have to work with many different businesses.

For parcel delivery, put very simply, instead of a van stopping and starting 100 times outside of 100 houses, the van would deliver those parcels to a Starship facility or wherever a delivery robot facility would be, drop off those 100 parcels. The van could go back to the bigger distribution hub and deliver another 100 parcels to another whole neighborhood or village or city. That's when the robots then take over. That's when you, as customers, would be notified on your mobile phone your delivery was ready.

We've seen huge efficiency gains, again, by going down that model. We're looking to eliminate missed deliveries. How many times have you been waiting around for delivery? How many times have you had your delivery thrown on your doorstep when you weren't there? Maybe you've had your delivery stolen. In the UK, we get delivery slips through our door all the time saying, "Hey. We missed you. We'll come back the next day," and they come back three days in a row at 2 p.m., and I'm still at work. We are looking to completely revolutionize and wipe out missed deliveries, with the parcel delivery firms that we've been working with, using these robots, and increasing convenience for the end customer and reducing costs for businesses, which of course is so important.

I've touched on this, but from the customer's side of things, simplicity is key. When you are introducing a new type of technology, you have to keep it simple. We're talking about robots on sidewalks. If it gets complicated, no one's going to know what to do and what to even think. You would receive a text message saying, "Your delivery is ready." You could choose a window of time within five minutes, very, very specific, and then the robot would ping you and say, "You could track your robot here." Just like we'd all track an Uber cab on our phone, you could track the robot. You then receive a second text message saying, "Knock, knock. Your delivery robot's outside." There's a link. You click it, and you unlock the robot. It's unique to you. No one else can unlock the robot.

One question that comes up a lot, "Surely I can steal that. I can just lift that up and run away with it and put it in my car." I know you're all thinking it. You could. I would be lying to you if I said it's impossible to steal the robot. However, we haven't had one theft yet in over 28,000 miles. It's met over five million people. It's traveled to over 60 cities, over 15 countries around the world. We haven't had a theft. It does have nine cameras around the front and back. It has two-way audio, so we can talk to you. It has sirens if you lift it up, so it's going to be an embarrassing thing running down the road with a screaming robot. It has tracking to the nearest inch, so of course we know exactly where you are, and the lid is locked. You could have a huge hammer and try and smash it open all for maybe some milk and eggs. It's really not worth it.

The robot technology. I've touched on the fact that we aim for 99% autonomous driving, not 100. We have obviously a very sophisticated obstacle detection system so it doesn't bump into things. It travels on sidewalks. Max speed, four miles an hour. It can carry around about 20 pounds. We're not going to be delivering flat screen TVs or sofas any time soon. Again, robots have a particular part of the delivery process that they can account for. However, it can actually deliver 90% of what we all order online on a regular basis.

Regulation and Laws

Now, regulation is very, very important. Is it actually legal to have a robot travel down the sidewalk? It totally depends where you are. Some countries it's legal without you having to change any laws. Some it's not. Robots traveling on the ground do have advantages over robots traveling in the air with air space regulations and things like that. We've found incredible positivity from the lawmakers that we have been discussing this with. Washington DC was the first place in the country, in the US, to legalize robots with the Personal Delivery Device Act of 2016 last year.

To date, we are legal in the whole state of Virginia, Idaho, soon to be Wisconsin, soon to be Florida, we hope. Legal in 10 to 15 other cities around the world. We have access to 150 million people with our delivery robots, which is amazing seeing as though this robot's only been around for a year on sidewalks. It's moving very, very quickly.

We have assisted with the laws. We actually have doubled the speed. We've doubled the weight of this robot to be able to grow into the industry. Again, it's an industry that's at the very start right now. There are other companies out there. We're not quite sure of their weights and sizes and speeds, but we've allowed for an element of grace there. We want other companies around. We want to build an industry here, so it's very, very important to note that.

Now, the final thing I'll say is one of the most surprising things that you may see. I'm sure none of you will believe me, but I can assure you this is true. The majority of people that see this robot on the sidewalk ignore it. I thought that everybody that saw this robot was going to be going crazy, crowding around it, taking pictures, trying to stop it, trying to mess with it. It doesn't happen. It simply doesn't happen. In over 28,000 miles of traveling, in many different locations, not just nice ones either, we have found the majority of people ignore it. They may look at it, but then they get on with their day. Most of them are looking at their phone, and they just don't care. It's amazing.

Social acceptance is absolutely key to the success of not just this company, but many other robotics companies, whether they're on the ground, whether they're in the air, whether they're on the road. It's something, again, lots of people overlook, but social acceptance is absolutely crucial.

Social acceptance is absolutely key to the success of not just this company, but many other robotics companies, whether they're on the ground, whether they're in the air, whether they're on the road. It's something, again, lots of people overlook, but social acceptance is absolutely crucial. It's a lot of the reason behind the design of this robot. I am biased. but I would say the robot looks very friendly. I would say it looks very cute. We've had children trying to feed it bananas before. We've had people pat it. I saw someone outside pat it like a dog, so people think it's like an animal. That's really important to us. It needs to give off that vibe, and people talk about emotional connections with robots. It sounds crazy, but it's true. People do actually get emotional connection with this robot, and we really care about that. We love that. We think it's really, really positive, and it assists with this business moving forward.

Transcript has been modified for clarity.


Speaker Bio: Henry Harris-Burland leads the Marketing and Communications team at Starship. After working at Rolls-Royce Motor Cars for a number of years, changing perceptions of an old fashioned brand, he moved into the exact opposite - delivery robots. Henry wants to change the way people perceive delivery.

This video comes from our inaugural shipping conference: Shippo Summit. Hundreds of customers, partners and industry experts from around the world joined us in San Francisco for a full day program of inspirational talks and dynamic conversations that explored leveraging shipping as a competitive advantage for ecommerce businesses.

Next post

Jun 16, 2017
Better Shipping for eBay Canada

Previous posts

Jun 03, 2017
Becoming an International Brand

May 31, 2017
Meet Shippo at IRCE 2017

Sign up for the Shippo Blog

Receive emails with news and announcements we post on this blog.