Amazon has taught me I don’t have to wait for my next two-pack of ravioli cutter stamps, so if you can’t get them to me in under two hours, I know someone else who can. It’s 2020, and customers want what they want, when they want it. According to a recent report, the global last-mile market is now expected to hit $55.2 billion by 2025, up from $30.2 billion today – and it’s no wonder. Amazon’s deep investments in delivery continue to fuel a surge in e-commerce; meanwhile, customer expectations and the entire supply chain have been completely upended.

The good news is that the more retailers invest in delivery, the more their e-commerce revenue grows. For businesses who’ve made supply chain a top priority, it’s huge validation.

So, where does that leave us in the race to the doorstep? Companies are throwing cash at everything, from drones to self-driving robots to crowdsourcing. Who’s got the best chance of success? How can each one lower costs, increase speed, and mitigate risks? Can they disrupt the industry without being, well… disruptive?

Drones: The Droids You’re Looking For?

Drones entered the mainstream about five years ago as a cool photography gadget. Thanks to falling prices, they were a hot item on every kid’s list this past Christmas, but they’ve also generated a lot of buzz about their potential applications for logistics.

In rural areas, drones have huge promise for parcel delivery. They’re already supplementing human workers in large warehouses – flying to far-flung corners to pick goods on high-up shelves. And they’re working out in the freight yard, too, helping to track and manage trucks, trailers, and containers.

Companies like UPS, Amazon, Google, and even Dominos are experimenting with drones in the last mile. One popular model uses a carrier van on the highway as a hub for an armada of drones that fly out of the back to deliver small parcels to nearby homes. It’s an impressive, futuristic version of hub and spoke. Can it work? At what cost?

Just like commercial aviation and the automobile, drones have major hurdles to navigate, especially when we think about how they’ll work at scale. We’ll need major regulatory oversight to address safety, noise, and privacy concerns. We’ll need to build control towers, write better algorithms, improve GPS, and figure out what to do about weather. But these challenges will likely all be worked out, given enough in time and investment.

Autonomous Vehicles and Robots: Bots with Brains.

What about autonomous cars and robots? Are they more viable in the near term?

McKinsey predicts autonomous vehicles could slash last-mile delivery costs in urban areas by as much as 40%. And companies from FedEx to Bosch have made bets on sidewalk delivery bots, deploying prototypes in San Francisco office parks, where they’re tightly controlled. Long-term, the potential is clear, and companies with the deep pockets to make early bets could save a lot of money in the long run.

Both autonomous cars and bots cost thousands of dollars per unit to manufacture, though, and depend heavily on human supervision and maintenance. When it comes to flexibility and scalability, is a sidewalk droid really that different from a truck? Both are rigid, asset-heavy systems that require a big capex investment upfront with even higher maintenance and upgrade costs over time.

Think of it this way: earlier this year, actor and comedian Tracy Morgan from Saturday Night Live and 30 Rock bought himself a sweet new supercar: a Bugatti. He forked out a cool $2 million for it. But later that day, he was sideswiped by a driver in a Honda CR-V. It was just a minor fender-bender, but it turns out fixing a scratch on a Bugatti costs more than the entire value of the car that hit him – somewhere to the tune of $32,000.

The truth is, even if I had the cash to buy a Bugatti, I could never afford to maintain it. Will a fleet of delivery droids be the same?

Whether it’s drones or robots or some other yet-to-emerge autonomous technology, asset-heavy logistics strategies will always suffer from the same Achilles heel: whether it’s changing wiper blades or switching from lithium batteries to solar – hardware is expensive. And if a new hardware solution can’t solve for the demands of flexibility in the last mile, there will still be a need for something that can.

This doesn’t even take into account all the regulatory hurdles, infrastructure dependencies, and real-world obstacles from bikes to baby strollers, pranksters to potholes, larcenists to labor unions.

But what folks aren’t talking about, and what I find most interesting, is the inherent limitations that come with any fixed-asset system.

Drones and robots may well be efficient, and hopefully one day safer. But what happens when a last-minute order comes in and the customer needs delivery now? How do you adjust a pre-planned droid route at the last minute – when the droid has already left the store?

Crowdsourcing: Using an Infrastructure That’s Already There

Robots may well be our future, but how do we solve the delivery challenges we have today? That’s where crowdsourcing comes in.

Crowdsourcing lets retailers leverage existing resources already on the road to make delivery faster, more efficient and more flexible. While others are making big bets on drones, crowdsourcing drivers are delivering things as varied as gigantic garden gnomes, temperature-controlled medicine that won’t be ready for pickup until 9 PM, and your online Home Depot order from this morning.

Some of the biggest brands in retail are investing in crowdsourcing. SMBs and Fortune 100 retailers team up with crowdsourced drivers to have them deliver everything from makeup to mattresses, paint to puppy food. Businesses across virtually every industry are solving today’s delivery challenges with an asset-light strategy that allows them to experiment and learn. They’re addressing delivery demand today, without making new capital investments or locking themselves into a futures bet with complex hardware systems. And most importantly, they’re not disrupting their existing supply chain in a way that can’t be undone without a huge cost if and when the autonomous tech winners begin to emerge over the next decade or two.

And that’s really the point. Retailers need optionality. Customers want to personalize their delivery for each and every purchase at the point of sale. A great customer experience means having a delivery solution for every customer delivery problem, whether you’re scheduling a sofa delivery on Sunday or sending a rescue inhaler right now. Making that work in the real world means having an arsenal of tools in your delivery toolkit. Crowdsourcing is a sustainable solution that ensures we’ll all be around to see what delivery looks like in the future.

Marc Gorlin is the Founder and CEO of Roadie, a crowdsourced delivery service that works with consumers, small businesses and national companies across virtually every industry to provide a faster, cheaper, more scalable solution for scheduled, same-day and urgent delivery. With over 150,000 verified drivers, Roadie covers 89% of U.S. households — the largest local same-day delivery footprint in the nation.

Follow