Food Delivery Apps Are Not Good Technology

What an unusually long escalator can tell us about innovation and business

Do We Need More Articles about Software?

Last summer, I expressed my disappointment with the prominence of software in our economy. It reads like the “atoms, not bits” argument you’ve seen in 100 other places. Still, it received more likes and comments than almost any other Simplify article, so I hope you got something out of it. I want to expand on my frustration with the software industry, and see if I can express myself with a more novel and interesting analysis.

I don’t always agree with “atoms, not bits.” I’ve praised Chat GPT for the hours of coding that it has already saved me. I like the fact that I can search Wiktionary to learn how “was” became the past tense of “be.” I like that Grammarly can find spelling, usage, and punctuation errors in my posts before they go live. Bits are cool, and a lot of human problems can be solved through bits. For example, I like to play solo board games. Most games aren’t optimized for solo play, and often see two types of disappointing one-player solutions. Some games just ask you to beat your high score. Others demand that the player manages a clunky automaton opponent through flowcharts and randomizers. One day, I’d like to play a game, scan the game state with my phone, and have some app select the opponent’s moves through some AI. Is this possible? I have no idea. If someone can program such a thing, though, I would enjoy it. Annoying solo modes in board games aren’t a major societal issue, sure, but it would represent an instance of software solving a problem. That’s nice. I’m for that.

The point is, though, that software can’t solve every problem. Take plastic waste. Researchers are developing plastic-digesting enzymes that would recycle the material and keep it out of natural habitats. If these come to fruition, they will require some software, and possibly even some AI. Ultimately, plastic waste can’t be removed on an LED screen. The stuff exists in the physical world, and we must fight it in that arena. In other words, some problems need atoms and others need bits. I see no contradiction here, nor do I intend to find “synthesis” between atom-dome and bit-dome. It’s a matter of different tools for different problems.

The Negative Balance Sheet Innovation

Let’s take food delivery. The fundamental problems here involve time and distance. Your home sits in a different place than the food does, and it takes time to move between those spaces. A possible solution to this problem involves delivery drones. I’m serious! Without such a drone, you (or a delivery driver) must drive to your place and back. This suffers from two issues. One, the food can’t take the shortest possible distance to your place, because there are buildings, sewage drains, and other undrivable stuff in the way. Second, roads contain cars, traffic signals, stop signs, speed limits, and other items that prevent maximum food-delivery speed. Drones offer a solution to both problems. It’s a bit of an excessive solution, sure. We’d probably find ourselves with additional issues in a drone-delivery world, and costs might exceed the benefits. Still, it’s an example of a technology that solves a problem. Can we say the same about delivery apps?

I’m not convinced that Grubhub, Uber Eats, or any similar firms create any new technology. These apps didn’t invent GPS, online payment processors, or mobile GUIs. I see no reason, then, that restaurants couldn’t have offered app-based delivery years before these apps did. We saw a few takers, of course. Pizza Hut has been delivering since before humans domesticated the horse. Yet, few firms took the bait and offered app-based delivery, despite access to all the relevant tech. Why was that?

There’s an old joke where an economist finds a $20 bill on the ground. He reasons that any $20 left on the ground would have been picked up by now, so he concludes that the $20 bill does not exist. I understand that my reasoning from the previous paragraph may suffer from the same fallacy. Maybe food delivery didn’t exist because no one has made it yet. Maybe Grubhub and Uber Eats filled a hole in the market that no one else had the vision or skill to capitalize on. That’s possible. It’s also possible that no one filled this market because it’s not a profitable one.

The finances point toward the latter. Enjoy these article titles:

• RECORD-SETTING QUARTER BRINGS GRUBHUB NO CLOSER TO PROFITABILITY

• DoorDash and Uber Eats Are Hot. They’re Still Not Making Money.

• Grubhub Shows That Food Delivery Alone Won’t Bring Profits

In short, these companies bleed money, with no clear path to profitability. These apps solve the aforementioned time and distance problems. They just don’t solve the “you gotta make more money than you spend” problem. I’m not making a normative argument here. There’s nothing wrong with unprofitable services. I don’t need to see a positive P&L statement from police and firefighters. Rather, I think that Grubhub et al don’t really innovate on the time or space front. Anyone could have made these apps before they did. Food delivery apps aren’t an innovation of bits or apps; they’re a financial innovation. Venture Capital cash has provided these firms the luxury of running a losing business model.

To understand this, let’s take a step back and remember why we have firms in the first place. Or, more accurately, let’s take a step up.

One Long Step for Mankind

At some point in life, you might find yourself attending a convention at Caesars Palace. If you do, you will search through the casino to find the convention halls. You’ll pass by the exorbitant Bacchanal Buffet and a comparatively cheap café plastered with photos of Giada De Laurentiis. There, you’ll find an escalator that leads up to the convention center. You’ll hop on the escalator and wait. And wait. And wait. And you’ll wonder, why the heck is this escalator so long? You won’t find an escalator that long at the company’s other hotels. There’s nothing like that at The Flamingo or Planet Hollywood, and certainly not at The Cromwell. Once you reach the top, you will probably continue with your day and forget about the ride. If not, you can pursue the Mystery of the Comically Long Escalator by heading to the pack of elevators. You might notice that that’s another story between the casino floor and the first set of conference rooms. The mystery seems: the escalator scales two floors, compared to the norm of a single floor. What’s in that skipped room, though? You can access it via the elevators, from which you will find a gaudy hallway that ends in a locked door. No, your room key doesn’t work. Behind the door, you see a lavish waiting room, where a few people shoot you a nasty look. You return to the convention halls and your brief stint as a detective comes to a close.

Or, at least, it would come to a close, if I didn’t tell you the answer: that’s the main corporate office. You won’t find anything similar at Harrah’s or The LINQ. Since all eight Caesars hotels on the Las Vegas Strip sit under the same corporate umbrella, they can save money. Businesses save money by combining operations. They don’t need eight sets of accountants or eight HR departments or eight corporate Salesforce licenses. A single oven can bake the muffins for your convention at Caesars Palace, the thematic coffee shops at Paris Las Vegas, and the employee break room at Horseshoe. They can purchase one set of trucks, not eight, to move all the necessary materials in and out of the hotels. The company can buy items in bulk, negotiating prices lower than any of the eight hotels could on their own. They can even save money on office rent by, you know, having everyone work one floor above the slot machines.

Combining operations also offer perks to the consumer. If you hit a hot streak at a lower-end casino like Harrah’s, the company may provide you with a lavish Caesars Palace suite. This wouldn’t occur if the properties didn’t belong to the same corporation. A convention organizer can hold the event in conference rooms at the Caesars Forum, with breakfast catering from Paris and lunch catering from Nobu. Afterward, all the guests can head back to their rooms in The LINQ

¹

and The Flamingo. To be clear, I’m not advocating for monopoly power! I can see obvious downsides from so much corporate consolidation. I’m just showing consumers can sometimes benefit from what companies like to call “synergy.”

I can’t say the same for DoorDash. From the supply side, drivers use their own cars. Compare this to fast-food pizza chains, where the company can deploy a shared fleet of vehicles. Different restaurants can’t share ingredients or loading docks. The consumer doesn’t see much benefit either. I can’t mix and match, ordering the entrée from my favorite Greek restaurant alongside a drink from my favorite coffee shop. It also doesn’t smooth out the delivery experience. A delivery service dedicated to a specific restaurant can assign different drivers to different zones such that every driver knows the area they’re delivering to. These employees would also have managers and relevant technology that makes the operation run smoothly. Instead of smooth operations, I’ve experienced:

• Drivers idling in front of restaurants for 30+ minutes, waiting for the restaurant to complete the order

• Orders arrive an 90 minutes late, due to a combination of idling and unfamiliarity with the territory

• Drivers take circles around the wrong apartment building, ignoring my calls

• Drivers take a 45-minute freeway route instead of a 10-minute one on surface streets

• My chosen items, despite appearing on the menu, don’t exist at the restaurant. Sometimes they call me and ask for a substitution, other times the order just gets canceled.

• Alternatively, the whole order gets canceled with no explanations

• I receive utensils despite declining them at every order. I now have a drawer filled with unwanted plastic utensils

• The order gets delivered to a nearby unit instead mine

• Substitution requests are ignored and sides aren’t included

  • One time an order came with two sides. I said I didn’t want one of the sides, so they didn’t deliver either

• The driver just eats the order. That’s my headcanon, at least, when it’s not delivered. To be honest, though, I kinda respect this one.

Sure, every service screws up sometimes, and there’s no company where everyone can attest to positive experiences. I am not arguing that Grubhub performs worse than the old delivery-by-phone system. I’m just showing that it doesn’t provide any technology that solves these issues. It’s as if the pizza delivery drone only drove above existing roads and followed all traffic laws. It amounts to putting the same issues into a different vessel.

Yes, these apps add convenience to people’s lives. I use it (as you can see from this article), and I enjoy getting restaurant food delivered to my house, even if it’s not always the food I ordered. Still, I think there was a reason that restaurant delivery remained scarce before these apps: no one has developed a profitable solution to the problems of time and space. The primary innovation was financial. Venture capital allows unprofitable business models to perpetuate.

A real, profitable solution to these problems would require investment in the material world. Here’s an idea. A company could construct a few multi-story facilities of ghost kitchens. Each facility would house a variety of restaurant-quality foods. As ghost kitchens, they would save on space, since they wouldn’t have to make any room for customers. On the supply side, these restaurants could exploit economies of scale. They’d own a fleet of cars, so multiple drivers could use the same car throughout the day. The company could save money by purchasing materials in bulk and sharing them throughout the ghost kitchens. This setup would also offer some perks for the consumer. Each delivery driver could specialize in a certain area, so they would know that Building 8 signifies the front of one apartment complex and the back of another. The vehicles could also specialize. Maybe one has special cupholders that can keep coffee drinks warm, while another has coolers for frozen yogurt and ice cream.

Could this work? I have no idea. But, like the pizza drone, it’s an attempt at creating a business that solves a problem. Yet, even if I could whip out a crystal ball and guarantee its profitability, it would require investment in the material world. For a variety of reasons, that just seems harder to come by.

My frustration, then, isn’t with software itself. Rather, it’s that software can only solve a limited range of problems. Of course, this holds true for almost anything. Earlier in the article, I mentioned plastic-digesting enzymes. These could address the prevalence of ocean microplastics, but plastic-digesting enzymes won’t cure Canker Sores or improve automobile safety. They certainly won’t help us determine the number of language families among indigenous North Americans. Yet, software alone seems to receive special status as the omni-solution.

I see this in my own line of work. Companies don’t need a 56th Tableau dashboard. They don’t need a data scientist that uses the latest machine learning model to solve a problem that could have been solved by the data analysis add-on in Excel. Companies need a coherent strategy, and they need analysts that can provide clear, interpretable reporting that supports that strategy. But that’s hard. On the other hand, it’s easy to hire a bunch of data people and keep them busy. I worry that our best and brightest (and me) spend too much time, too much energy, and too much VC money on non-solutions to real problems.

1. The actual, honest-to-God, official name is “The LINQ Hotel + Experience.”