Source: Jeff Sandquist via Wikimedia.

Amazon.com (AMZN -1.68%) has long pursued grocery delivery, but as a recent change to its AmazonFresh program demonstrates, it has yet to crack the market's proverbial nut.

AmazonFresh hits a speedbump
On-demand grocery delivery has always been something of an albatross. During the Dot-com boom, start-ups like Webvan and HomeGrocer.com tried to harness the power and convenience of the Web to create early grocery-delivery services. Predictably, the onset of the tech bust sent both concepts the way of the dodo bird. But with the grocery industry driving sales north of $600 billion annually, the opportunity is too tempting for a company like Amazon to resist.

Amazon resuscitated this business model when it introduced a beta version of AmazonFresh in Seattle in 2007. Since then, it's expanded the program to include parts of Los Angeles, San Francisco, Philadelphia, and New York City. Although it has always charged a fee in LA, Amazon has included a free promotional period in most other cities in which it operated. But this month, Amazon began notifying AmazonFresh users that the delivery service would cost $299 for an annual subscription fee, triggering a wave of complaints.

Grocery delivery's economics must evolve
As a company with a world-class distribution system, Amazon's move to charge for AmazonFresh service underscores the challenging economics that accompany the service.

One problem is the so-called "last mile" problem. In its normal e-commerce business, Amazon outsources its delivery services to third-party providers, such as UPS or FedEx. Because grocery delivery requires vehicles fitted with refrigerated compartments, however, Amazon is forced to operate its own delivery trucks, which adds considerable expense. The good news is that Amazon enjoys a large and diversified revenue base from which it can support the pilot program until more cost-effective delivery methods (namely, drones) prove viable.

It also appears as if the distribution systems from grocery retailers to Amazon, or to competitors like Instacart or Alphabet's Google Express, aren't scalable in their current form. AmazonFresh and Instacart claim that they can combine goods from third-party retailers into a single order -- even from small, city-specific local stores. But because tiny retailers aren't large enough to connect with Amazon's distribution network, someone still has to go physically pick up a jar of, say, artisanal cupcakes from across Manhattan to fill my hypothetical order. I trust we can all agree that isn't efficient.

Source: Amazon.

Giving away $1.00 for $0.85
Recently, comments from legendary venture capitalist Bill Gurley about AmazonFresh's competitors, including privately held Instacart and Alphabet's pilot program Google Express, addressed the hurdles that persist in keeping this seemingly appealing business model confined to a niche service: "It’s like the old adage, [when you’re] handing out dollars for 85 cents, you can go on [infinitely],” Gurley said. “Chosen unicorns are being given hundreds of millions of dollars, but you have to ask how much margin is there. The unit economics ... would be very difficult, I’d think."

Although the financial side of many tech firms' businesses remain a black box, it's clear that key structural barriers still exist that prevent grocery-delivery services from gaining mainstream adoption. Automation of the supply chain, both on the fulfillment and delivery side, will go a long way to alleviate some of these pain points. But with self-driving cars and drones still years away, the average consumer or tech investor shouldn't get hung up on the potential that could come from disrupting the grocery delivery market today.