But because they have an insurmountable most (not earned via any sheer technical breakthrough, just through cunningness and being ruthless) they don't need to care if they have too many engineers,whether these engineers are any good, or if their nodejs ec2 instances serve 100 queries a minute. So they just keep it like that.

Companies like IBM needed decades to get into such a rot, modern VC funding and cloud computing have just accelerated this phenomenon.

We've been told that evil VCs subsidize service to run at below cost to drive competitors out of business. Uber is often accused of this. Why aren't we seeing that phenomenon here? And Uber is literally one of the competitors.

There are some other possible reasons tho, like food actually not being as "commodity"-like as we think it is (like rides are)

The other hypothesis is that delivery is simply expensive. The driver has to drive to the pickup place, find parking, stand in line potentially for awhile, drive to the delivery, park again, walk to the delivery location to drop off the item. Most of those things an ordinary uber fare would not have to deal with.

Reminder that Adobe, Apple, Google, Intel, Intuit, Pixar, Lucasfilm and eBay all colluded to keep engineer compensation below market value[1].

[1] https://en.wikipedia.org/wiki/High-Tech_Employee_Antitrust_L...

A claim that all of these have formed a cartel requires strong evidence.

And 2 of those are potentially combining.

Looks like a split amongst 4 right now, as far as the big ones go. This doesn't include the smaller startups as well.

If you went to a VC with a model for a food service delivery taking less than 30% of the transaction, they'd more than likely pass. 30% is the bar that's been set, and your unit economics have to be as good or better than the existing players. They're not keen on investing in "just a cheaper version of something that already exists." -- If you're marketing yourself as "cheaper," it has to because some innovation has eliminated more cost out of the transaction than you have eliminated in net revenue.

And tbh this should happen. UBER had to spend billions on marketing and getting their model "right". But now it is clear that such platform is essential for modern city. At some point UBER will have to get this money back with ride fees. Next competitor that can replicate UX of UBER, but raises less capital will be very well positioned to offer competitive pricing. And switching between these platforms is much easier compared to something like Facebook/Apple.

Grub-less et.al are as much about ordering and checkout as they are about discovery. I can rank and view all restaurants offering Sushi in my delivery radius. This is what makes them powerful.

Uber et.al are a marketplace. You are absolutely right in that they have zero switching costs but they have a two sided pool of supply (drivers) and riders (demand). This is hard to replicate.

I suspect most customers are either starting looking for a restaurant they know by name, or one near them to ensure maximum freshness, using Google Maps or Yelp or the like. Only when they've already chosen a restaurant do we transition into the delivery app flow.

I'm sure a lot of money is being spent for chain restaurants to advertise "order via Postmates" or "exclusively on GrubHub"-- again bypassing them as a discovery system.

There is an argument to be made, that if I just want a pizza, then my major decision-maker will be price.

Your moat argument is the same that had people invest in uber 10 years ago, which we are seeing is failing in regions like EU where Uber is getting lambasted.

Of course, you'd have to be rich in order to have done that in the first place (an "accredited investor" as defined by the SEC), and this ignores the other investments you made in 2010 that went bust.

Later investors got screwed by the poor IPO performance, especially those that bought, hard, into the "moat" argument. Hell, softbank is possibly underwater on its investment.

Visa, Amazon, Apple, Google, or Paypal sort of provide that function, but being able to shop many restaurants with a single identity has its benefits. Something that wouldnt be fixed just by restaurants with nice websites.

In the long run, I honestly think kitchen-only setups optimized for efficiency & food safety will grab a big share of the market. They can experiment with vertical integration, and avoid all the hassles & limitations of sit-down dining. They will be able to more, better, faster and cheaper. It might be hard to copy & paste the entire thing from one city to another, but the core functions should be quite portable.

It would be crazy smart for someone to standardize an interface (e.g. REST API), let restaurants implement glue to their systems, and then allow a plurality of delivery services to consume and order through the standard interface.

It needs to be a protocol like email or http, not another company with a proprietary aggregator.

For delivery it's harder because you actually need someone to deliver the food and the logistics are really hard (they have to get there when the food is ready but not too late and not too early, etc.). But... there are like 80 restaurants in a 10 minute walk radius from me, so I can just deal with pickup for now.

The best part is actually just having a variety to order from. Before these delivery services, there were only a couple of pizza places and a couple of Chinese restaurants and nothing else delivered.

For a lot of restaurants, it never made business sense to deliver, and still wouldn't. But when the delivery service is spread out over many restaurants, it works rather well. That's good for both the businesses and the customers, despite the friction added by the middlemen. It's a real win for people who couldn't as easily go out to eat, due to health or disability or whatever, but also for everyone else.

In such a world, Some model of Ghost kitchens will thrive anyway, not small restaurants.

Why ?

The area a kitchen serves will become much larger - distance matters less when you don't need to go pick the food. And even when food freshness matters, there are some solutions like parcooking and finishing the cooking in ovens inside delivery vans.

The type of food a kitchen serves becomes much more varied - today your Chinese restaurant has one kitchen, and the Mexican place has a different one. But with delivery-only, a single kitchen can open many virtual restaurants.

This also helps with delivery - a single family who likes different food, can order from the same kitchen, and get the order delivered in one go.

And this is now an internet business - since customers prefer all the restaurants in a single app, single search - and those add new complexities.

And usually, small business(in aggregate) often had the benefit of agility over large businesses. But a virtual restaurant is very agile, conducting business experiments is rather cheap.

And personal service and deep relationships don't exist in food delivery - and that was a weakness for big businesses over small ones.

So long term i don't think delivery a good place for restaurants to be.

Which only seems compelling in a future in which people are literally unable to cook for themselves.

I have worked with these large kitchens that can make different types of cuisines depending on your want, but they all turn out mediocre.

My family completely stopped ordering out during this pandemic because most restaurants have lost the key employees who make their food special, there is no atmosphere, and GrubHub et al add no value.

What about the delivery part of the problem? Everyone does realize that Seamless etc. actually provide the delivery staff for many of the restaurants on the platform, right?

That's only half the battle though, food delivery marketplaces like Grubhub are very entrenched and changing consumer behavior back to using the restaurant's website takes time and effort. Marketplaces also spend a ton on marketing to make sure consumers go through them and not directly via the restaurant's website. It's an uphill battle for the restaurants.

As an example, most people buy from amazon even though other retailers have very good websites, without inline advertisements and many times with lower prices.

another interesting example is tow truck services.

Some retailers have good websites (most don't), but I've always found issues like: slow shipping, having to create an account, spam from the retailer, poor customer service, etc.