Iceland continues to act as a proving ground for the use of drones in package-delivery services.

Last year, the Israeli company Flytrex began experimenting with drones in Reykjavik. It launched hundreds of flights that bypassed traffic congestion to deliver food and consumer goods to an outlying residential neighborhood. The tests began with a single delivery point and subsequently expanded to between 40 and 50 locations around the city — half public places and half private backyards, according to CEO Yariv Bash.

Bash rates the initiative a success, even though the flights were temporarily halted by Iceland’s harsh winter. (In a few years, he said, the craft will be able to operate during that season as well.) Initially, the destinations were almost all public places, with pilots attempting a handful of backyards “just to show the local regulator that it can be done safely.” Now the drones have been approved for more drops on private property.

Flytrex has aspirations reaching well beyond Iceland. According to Bash, its drone program is one of just ten to be accepted into the U.S. Federal Aviation Administration’s Unmanned Aircraft System (UAS) Integration Pilot Program. The effort was launched to speed up the safe incorporation of drones into the nation’s air-traffic system.

Bash expects Flytrex to begin limited operations in the U.S. within the next few months, with food deliveries in Holly Springs, North Carolina, a suburb of Raleigh. The company will focus on delivering to backyards, where its drones offer “the best user experience,” he says. Apartment buildings pose a problem because they require the receiver to exit the building or access the rooftop.

There’s no question that a well-designed drone can drop a package into the receiver’s yard with pinpoint accuracy. (Instead of landing, Flytrex’s models hover between 50 and 100 feet off the ground, and lower packages by wire.) But are drones economically feasible for that purpose?

Bash thinks so. He argues that drones are significantly cheaper than the alternatives, such as autos, scooters and other labor-intensive options. All it took to launch the drone program in Iceland was two delivery persons and two days’ worth of training, so that they could pilot the craft beyond operators’ line of sight. (That capability is still banned or severely limited by the FAA in the U.S.)

Even doing one delivery at a time, the drones have proved themselves economically, Bash claims. He expects the numbers to grow even stronger “once we start doing multiple deliveries [at a time] and have greater bandwidth.” Current drone technology prevents that from occurring now, “but we’re almost there.”

Other concerns about the widespread use of delivery drones have centered on safety, noise and privacy. By using a wire to lower the package to the ground, Flytrex ensures that the craft remains well above the recipient. (Pulling on the wire automatically detaches it from the drone.) And because it hovers high off the ground, noise from the rotors is minimized.

Privacy is another concern altogether — one that probably won’t be resolved until drone deliveries are well underway, and regulations put into place. By remaining at a fixed height above residences, drones theoretically don’t risk unwanted forays into private property. But certain angry individuals, believing them to be instruments of surveillance, could attempt to damage or destroy the craft.

Bash discounts the likelihood of such an occurrence. “A drone hovering at 100 feet is not an easy target,” he says, adding that while Flytrex’s drones haven’t experienced any attacks to date, a few people have attempted to pull the wires.

He sees incremental progress being made toward allowing for full operation beyond operator’s line of sight in the U.S. “FAA is highly motivated to make this work, but they want to have a look at the larger picture,” he says, describing the agency’s approach as “step by step.”

Bash had previously predicted that drones could begin operating in wide-scale service within the U.S. as early as 2019. Now, he’s expecting regular services to commence around the middle of 2020, following the end of FAA-approved pilot programs.

As in the case of Flytrex, drone operators at the outset are likely to confine service to suburban locations, home to two-thirds of the U.S. population. (Bash describes that demographic as consisting of more than 60 million backyards.) The use of drones in crowded urban areas is likely to take longer, while designers work out the logistics of serving multi-unit residential buildings. The solution, Bash says, might lie with a combination of technologies — say, a drone delivering to the designated building, and a robot carrying the package inside and possibly to the receiver’s door.

In pushing for the commercial use of drones, Flytrex has learned some important lessons along the way. One is that landing at destination is a bad idea.

“I don’t want anyone near my $5,000 drone,” Bash says. “It’s not good for the drone or the person.” Adoption of the hovering model solved problems of safety, security and noise.

Drones still have a ways to go before questions of practicality and economic viability are fully solved. But Flytrex believes the technology is set to play a major role in the last-mile segment of the burgeoning e-commerce supply chain.

“Everyone’s pushing really hard for commercial drones,” Bash says, adding that the last-mile challenge to date “has been a pain in the neck for retailers.”