Hacker News Reader

Comment by seanmcdirmid | Hacker News Reader

Food delivery also isn’t very practical in developed country markets without a huge increase in productivity vs labor costs. It’s just too expensive compared to a country like China where it is cheap enough to be ubiquitous.

Automation could be disruptive here…someday in the form of food delivery land and/or air drones. It really is the same with ride share: it just isn’t economical considering how much human labor is worth, self driving vehicles are more than likely required to make it economical enough to be ubiquitous.

View on HN

Food delivery is practical, mainstream and has been for 40+ years in the West.

It works great "inhouse". They tend to know all their regular customers and area and know exactly how to get there, where to park, which bell to press etc. They will also know if there is bad traffic or an accident closing a major road and adjust wait times and order frequency accordingly.

The person that runs the delivery joint can also wait until he knows his/her driver will be making his way back, and only start cooking then, so the next order doesn't get cold waiting for a pickup.

It all goes to shit when you put a tech platform which has absolutely no real interaction with the kitchen and outsource the riders entirely, with virtually no quality control and no local knowledge. Plus they also want to take 30-50% commission for the privilege.

Ubers margins are comically bad. Uber Eats did $13bn of revenue and they made $30m of EBITDA - 0.2%. The credit card networks are probably making 10x the money that Uber on each order. And that EBITDA will be with them pulling out all the stops now they have realised they need to get profitable fast.

>They tend to know all their regular customers and area and know exactly how to get there

>They will also know if there is bad traffic or an accident closing a major road and adjust wait times and order frequency accordingly.

irrelevant in the age of GPS navigation with live traffic (ie. google maps)

>where to park, which bell to press etc.

while I don't doubt delivery drivers accrue such knowledge over time, I doubt that the turnover is low enough for this to actually make a difference. It's certainly not enough to make such a service unviable.

>The person that runs the delivery joint can also wait until he knows his/her driver will be making his way back, and only start cooking then, so the next order doesn't get cold waiting for a pickup.

This doesn't seem very hard to do. Restaurants can let the delivery platform know how long each dish to prepare, and the delivery platform can tell them ahead of time when to prepare the food (after factoring in driver availability and distance).

In urban areas (where most of the demand comes from), Google Maps does not provide the detail required to get this right. If you are delivering to an apartment block you often can't just park on the street outside, you'll get ticketed straight away. This can make an enormous amount of difference, I've had uber eats take 10-15 minutes to find the (quite easy) building I live in, vs seconds from local takeaways who know the area.

This is one of the reasons uber eats often delivers horrendously cold food (with no refund option these days). If you got cold food from a local joint with inhouse delivery they'd usually refund or redeliver for free, because they know it was their fault. They can't offer that through these platforms as it is rarely their fault and even if it was they don't get an option to interface with the customer like that.

On the final point - firstly, it's not as simple as that as not 100% of orders are coming through uber eats, so they don't really know exactly how busy the kitchen is. Secondly, uber was so desperate to grow market share they would push kitchens to take orders regardless of capacity to serve them. The only option AFIAK was to shut off Uber orders entirely, there was no "take 4 orders per 30 minutes max" option (maybe they have added that now).

> irrelevant in the age of GPS navigation with live traffic (ie. google maps)

The confused phone calls I get from delivery drivers just about every time I order begs to differ.

And that's assuming they bother to call rather than just leaving my order with the basement apartment or one of my neighbors or in the same address only on Northeast [name] street 8 blocks away rather than Northwest [name] street.

These drivers churn so much they never learn their way out of routine dumb mistakes like that.

Or worse, they don't see the negative feedback - meaning even if they stick around long enough to handle your second order, they still won't know any better.

This is the actual problem with the app model. No one is on the line for your order being late, cold or wrong. The delivery guy doesn't care - he's the most transient and precarious labor possible and he only needs to get a certain percentage right to keep working. The restaurant, traditionally, tried very hard to keep their customers happy, and would send the (in-house) driver back with forgotten items or free stuff if something went wrong. Now they are plugged into a third-party system they don't have any confidence in or understanding of. They do their best but they can't predict demand, be sure when the driver will turn up, or be sure that complaints are legitimate. Meanwhile the app operator doesn't care that much about your order either. Getting every single order right would massively increase their costs while hardly adding much revenue through brand loyalty. Signing up new restaurants or rolling out to new cities is always going to be a more promising strategy.

Meanwhile, you are hungry and tired and you just spent $50 on $30 worth of takeout. You really really want someone to care about your dinner as much as you do. But next week when you're hungry and tired again, won't you roll the dice again and hope they get it right this time?

To make it worse, I have gate code to enter my complex - I have succinct, exact instructions including the gate code and how to navigate to my building. 90% of the time they never even read the instructions, call to ask for a gate code and then take 10 min to find my place.

I can see that happening pretty easy. Are they looking at their phone, reading instructions as they arrive? No, they're driving. They get to the gate, need a code, hit 'call'. I would do that.

If I lived in a Chinese Puzzle Box I would have to get used to people wanting help to open it. Because they don't actually give a damn about the puzzle box. They just want to drive straight there.

In places like Atlanta, gated complexes are common enough that it makes sense to read the directions, maybe while at the restaurant grabbing the food. Even better, one of the 3000 developers could make the app prominently display the gate instructions when they approach.

I ran into similar issues many times. I don't even have a gate code, but finding the building can be confusing. They do not read the instructions.

> Food delivery... works great "inhouse".

but

> Plus they also want to take 30-50% commission for the privilege.

If food delivery works great in house then why do restaurants pay Uber so much? I'd wager employing drivers is impractical for a lot of restaurants.

Well, McDonalds was a crazy amount of UberEats business. I heard as much as 50% of it. McDonalds is now in the UK offering their own inhouse delivery service. Clearly they've let Uber do all the hard work proving it works and now want that commission back.

At least in my area, McDelivery is provided by Uber Eats.

All this seems to mean in practice is I get a crappy Uber Eats delivery but have to order it through the McDonalds app, which is somehow even worse than Uber Eats.

I've just looked at the McDonalds app, and it says my local McDonalds is 24/7 and lets me order to collect from there. I know for a fact that it's shut right now, and the official store locator also says it's shut.

I stick to Deliveroo and Just Eat where possible because their apps and service tend to be pretty decent.

“All of your competitors pay us 30% so they can order food from their iPhone, pay us the same amount so they won’t take your customers away”

Doordash would offer a restaurant before partnering with them by calling in the orders so they could have sales numbers to pitch to the restaurant owner when they want to take the 30%.

It doesn't work great in house. It works great with economies of scale. I don't know the history of pizza delivery, so this is speculation, but I imagine that pizzas travel well, and people became accustomed to ordering them for delivery. So a driver can load up on pizzas, make a whole bunch of deliveries (and people are generally satisfied with the quality), rinse, and repeat. That's economical. Delivering a handful of orders at a time (maybe even 1) is obviously a terrible value proposition. But if it's incremental revenue, and covers your cost of goods sold, you may as well let Uber lose money to do it.

In the '90s when you wanted to order delivery, you'd open a phone book and pick a restaurant. Some would pay extra for a big, extra-visible listing in that phone book.

Uber, Doordash, etc get away with charging such premiums not because delivery is so costly but because they've made themselves into a new phone book. Hungry customers open the app to look for a restaurant, and the restaurant wants to be listed there.

However that's not really a win for the restaurant. It might not increase their business, just stop them from losing it. If the app didn't exist at all, those customers would find them another, less convenient way. But because the app exists, the customers stopped using those ways.

That's a good point. Once DoorDash hits critical mass, I agree it's as much a shopping app as delivery service. I also think that early in a market they benefit from making delivery possible for restaurants that otherwise wouldn't have it.

Only a few high volume high margin restaurants are able to do in house delivery economically. Typically pizza joints where delivery is already baked in to the price as that is their primary sales funnel.

>Only a few high volume high margin restaurants are able to do in house delivery economically. Typically pizza joints where delivery is already baked in to the price as that is their primary sales funnel.

That's not true where I live (obviously, YMMV). Granted, I live in a very urban area (NYC) and most deliveries are done on bicycle, as the distances are shorter and it's much faster for the delivery folks.

Most restaurants near me have had delivery services for decades. Depending on the restaurant, they might have dedicated delivery folks, but others would just send a busboy as needed.

In fact, the big problem used to be all the restaurants leaving their delivery menus in building lobbies or even slipping them under the doors of non-customers.

Now the problem is that many of these restaurants are going out of business because the fees on the "delivery" apps eat up their entire gross profit margins.

NYC is on a whole other scale than just about any other city in North America. The density of NYC changes the economics of just about anything drastically.

>NYC is on a whole other scale than just about any other city in North America. The density of NYC changes the economics of just about anything drastically.

Absolutely. Which is why I qualified my statement with "YMMV."

That said, I think one aspect is broadly applicable: Most low to mid-price restaurants (not chain restaurants) operate on pretty thin margins. And the 12-30% (just for "listing" on their website) can often eat all the profits from the business.

And that doesn't include the "service" fees and delivery charges either. Making the whole enterprise a bad deal for everyone except the UberEats/DoorDash/GrubHub/etc.

Edit: Fixed typo.

I'm unconvinced. Restaurants are not "forced" to eat those service fees, they can up their prices on delivery apps to compensate.

Does that make it a bad deal for the consumer? Maybe, but lots of people seem to be willing to pay the premium for the experience and convenience of ordering through the app.

Locally I've seen plenty of restaurants have different prices on UberEats than online.

>I'm unconvinced. Restaurants are not "forced" to eat those service fees, they can up their prices on delivery apps to compensate.

I'm talking about service fees to the customer, not the restaurant. The restaurant is already paying 12-30% on every single order. That includes phone orders via the fake websites these services set up with a VOIP number (not the restaurants' own numbers) that forwards to the restaurant.

As for upping the menu prices, IIUC, these services (GrubHub definitely, the others I'm sure do the same) require restaurants to charge the same prices as their eat-in menu.

You're welcome to disagree as to whether or not it's a good or bad deal for anyone, but I've seen (and had it confirmed by several former restaurant owners) restaurants go out of business because they, can't make a profit via the delivery services. Even ones that had been doing just fine for decades.

>Locally I've seen plenty of restaurants have different prices on UberEats than online.

Fair enough. I wouldn't use UberAnything even if it was free, so I wouldn't know. But I am certain about GrubHub/Seamless.

I expect you're still "unconvinced." Which is no skin off my nose. I'm loathe to tell anyone what to think, do or say, as it's neither my business nor my concern.

Uber eats has to drive to the restaurant each time. The pizza guy can gab 5 at once.

I live in suburbia and gave up on food delivery years ago. The drivers would regularly get lost. I would provide directions on how to find my apartment building and unit number. They would arrive at the complex, and still get lost, having not read the instructions. Some drivers would call, say "I'm in the parking lot", and expect me to find them (there's 8 different buildings.)

It's cheaper and more convenient to pick the food up myself.

>Food delivery also isn’t very practical in developed country markets without a huge increase in productivity vs labor costs.

But pizza and chinese delivery was a thing in the US long before delivery apps were a thing. Clearly it was practical at some small scale in the US. I don't see any reason why it can't be expanded to other restaurants (eg. by sharing delivery drivers so a restaurant doesn't have to have an entire FTE on payroll).

It wasn't bad before the gig-economy middlemen stepped in, in highschool i delievered pizzas and pocketed something like $2 + tips for each delivery (this was 20 something years ago )

it scaled pretty well, I would just take 3-4 orders per-ride depending on distance ended up with a decent chunk of change just from the weekends, and there was a decent amount of downtime and socializing waiting for pizzas to be cooked

with the gig stuff drivers are shuttleing between multiple resturants non-stop in ill-defined areas to random customers and probably walking away with less cash it's a stupid system

Ya it was a thing before when labor costs were cheaper. Will It remain a thing now? We still get delivery, some Chinese places don’t even bother with a dining room, but the costs keep increasing such that I don’t see it lasting for much longer.

> Food delivery also isn’t very practical in developed country markets without a huge increase in productivity vs labor costs

What on earth are you talking about. Pizza and Chinese restaurants have been profitably delivering food for decades.

The UberEats model in particular may be unprofitable, but there are absolutely profitable models that have existed for far longer than I’ve been alive.

It seems you’re right because all the major companies in the space are running at huge losses and not just due to an intentional choice to scale rapidly. But it also seems very surprising this is true.

Uber gets away with charging one side of the market 30% of food prices and charging the other side of the market a delivery fee, a service fee and a tip. This typically comes out to 50% of food costs if you use an average tip.

Here in LA in some areas, little trucks drive around on sidewalks delivering food driving remotely by an assumably low-paid driver.

Reply on news.ycombinator.com