Quick Grocery Delivery is impossible and undesirable, but there is another way
Aaron Gordon’s latest bit in Vice, on the (exceedingly false) promise of 15-minute grocery delivery, is really good:
These businesses all aim to fulfill the same purpose, and they all seem almost too good to be true. They offer a similar variety of convenience store items, promise the same basic delivery times, and operate in a similar array of high-income neighborhoods in cities around the world.
It is worth, I think, going a bit further than Gordon’s sources in the skepticism: it is not just that the algorithmically-generated delivery has incredibly tight margins, that the convenience factor is overrated, or that the wash of “reducing food waste” is “slightly irresponsible.”
It’s also mathematically impossible in the status quo United States. Customers are either too far apart in not-dense-enough locations, or too diverse in their interests if they are in a dense place.
Ingredients for Success
Gordon’s focus is on Jokr, the 15-minute-delivery startup in New York City that has set up shop already in Brazil, Colombia, Peru, Mexico, Poland, and Austria. As he notes, they are getting by on investor money.
Jokr is one of many similar startups that gets tracked by Matt Newberg’s HNGRY news service that focuses on the food/fulfillment/delivery industry. In my time reading HNGRY, I’ve only seen one service that actually smells and seems successful by traditional business metrics (i.e. turning a profit): Picnic.
Picnic’s (claimed) difference is not to use AI to predict orders, it’s using AI to sort logistics. A company blog post, “The Milkman Van for the 21st Century”, lays it out:
one of the ways that we can improve vehicle utilization is by automating the process of planning vehicles to delivery trips. Until now, this is a manual process that relies on the puzzling skills of the operations manager at the distribution hub to make an efficient schedule. Since we started doing deliveries in the mornings, we see that some of the older vehicles are not able to drive all back-to-back trips, especially if some of those trips cover longer distances.
Part of this is straightforward 21st-century logistics, but part of it is also, well, the Netherlands. Cities have dense housing (can serve many customers in one short trip) and a monoculture of choice (yes, there are immigrant groups in the netherlands. No, they are not the target audience here). There aren’t corner groceries or late-night markets in the same way the US or even Germany or France have. My experiences are eight freaking years out of date, but in Utrecht there was Albert Heijn and a handful of ethnic markets. There was a line down the block and giddy excitement when a German restaurant, of all things, opened. The situation, judging from a quick Google, has not improved:
That gives Picnic room to work: there is predictability and density, which frankly stacks the deck for artificial intelligence versus human choice.
Depth of Choice in the USA
The predictability + density duo is historically what gave rise to the United States’ most beloved food logistics hub: the bodega. Corner stores worked, in the early 20th century, in large part because they were the supply for what locals needed and weren’t in supply at the supermarket. It has by-and-large disappeared as a new use (though extant shops hang on) due to three things:
Smaller households and single-family conversion lowering number of customers within walking distance
Rise of residential-only zoning, even in moderately dense areas
Diversity of populations and of taste, meaning you can’t just stock for “one” ethnicity in high-end neighborhoods with lots of new residential development
In short, there are fewer people and fewer opportunities. Except for the places that have lots of people and lots of opportunities, but then there isn’t enough shelf space for folks looking for both mirin and stroopwafels.
There are a few attempts to redo the corner store in the United States. Foxtrot is probably the most notable, as it’s pushing to expand its “Upscale 7-11” concept. They really rely on high-markup options (i.e. booze, charcuterie) and the one I remember passing by, on Division in Chicago, had a “we need to grab something for the party your lawfirm colleague is hosting” vibe. It seemed like a tough competition next to the mini-Target next door, but they both are in business a few years later so hey, what do I know?
But this goes into a second, more obvious, issue in the US: how dang stratified urban living is. There are very few places dense enough to support 15-minute delivery in this country. Those places are either full of wealthy people who can pay a premium (in which case it’s not a delivery service, it’s a concierge) or impoverished people who the delivery system is point-blank not designed for.
The use case given of someone needing milk - right this instant! - is illustrative here. A family with kids can afford to get a giant jug of milk once a week, or they can’t afford to. There’s a vanishingly small middle ground. People who need oat milk for their lattes aren’t going to get milk delivered, they’re just going to go to a coffeeshop, which is probably the one US food access point in great supply.
It’s near impossible to do the back-end logistics like Picnic does in the Netherlands because the United States is not built like the Netherlands, and even where it is - does not have the demographic profile of the Netherlands. Which, as Gordon points out, means that you have to really really know what customers want. Nobody’s been able to do that besides Amazon. Are you Amazon?
Delivery Without Technology
In his article, Gordon wraps up by quoting Arizona State University planning professor David King on what makes cities good:
"What we like about cities are the frictions. It's the frictions that make cities economically robust. The frictions that make cities fun to be in... Those are the types of things that people like to have. It's not about just having your oat milk delivered in ten minutes.”
It’s one thing - a very valuable thing - to point out that quick grocery delivery is at best irrelevant and at worst directly conflicting with what makes cities good. It is another thing to point out that quick grocery delivery is impossible, or at least very unlikely barring mass data scraping at a scale that competes with Amazon. It will take billions of dollars and a moonshot to get quick delivery to work. Or, you could go the Netherlands route: densification and easier, safer, local logistics to smooth out the demographic wrinkles that make American cities so American.
I referenced this before, with ghost kitchens: if you can support a legitimately attractive use, particularly one that requires labor (such as kitchen work, or last-mile logistic hubs), and support it with land use, housing, and transportation decisions...then you really have a way to get people what they want, quickly. This also requires re-designing the American city, thinking in terms of logistics instead of investments.
You can probably see the issue here: relying on venture capital to put logistics in front of investments completely misunderstands the purpose of venture capital. Quick delivery is soaking up money because people want to capture data on high-value customers, not because any of these companies are founded to make quick delivery a social good.
You could imagine a socially useful example in here somewhere, something sort of like the window delivery system that elderly people use in Istanbul. There is a system that gets young, able-bodied, people who have time to make a buck live and work next to people who need their help. But whatever system that is, it doesn’t exist in the United States.