When I order delivery from a restaurant, an employee (maybe the owner’s kid) comes and delivers the food directly to my house when it’s done. It‘s generally faster than the app model where you need to wait for the driver to go to the restaurant to pick up the food, then come to your house while possibly stopping at other restaurants and/or houses on the way.
I could see how the restaurant delivery would be worse if that particular restaurant is bad at delivering. Also, the app model does seem like it should be cheaper at scale by using the deliverer’s labor more efficiently.
The restaurant also has a more direct incentive to actually deliver.
One Friday night here in Sydney, it was raining like crazy - so we ordered in some food from one of the online groups. 2 hours later and multiple talks with support and they were still telling us stories about how it's nearly there.
We eventually called the restaurant directly and the food had been sitting ready for 90 minutes on the counter. The service just had all their gig-economy-staff decide this isn't the gig for them in bad weather, so no one was available to pick it up and deliver it. They said we could pick it up ourselves but we didn't really feel like old food at that point so we cancelled and drove out somewhere else. Being lied to was the most annoying part.
I was picking up sandwiches from a local shop and watched 3 doordash/Uber Eats drivers come right after the other to pick up the same order, but they all cancelled it because the restaurant didn't have multiple deliveries for them to pick up at the same time- ie- they were dropping off multiple houses each run. It made me realize why the food was cold last time we ordered from them! From now on I just go and pick up myself.
> the app model where you need to wait for the driver to go to the restaurant to pick up the food, then come to your house while possibly stopping at other restaurants and/or houses on the way
This is a weird set of complaints.
1. Usually, there would be some lag time between when you order the food, and when the restaurant finishes making it. This would overlap with the time required for the driver to go to the restaurant and pick up the food; the app model isn't actually introducing a delay there.
2. Restaurants that do delivery themselves also use the same driver to deliver multiple orders at once. You're definitely better off with a dedicated driver, but the restaurant is much worse off. Again, this would not appear to be related to the app model. We order pizza and pick it up ourselves, because if the pizza place delivers it, it arrives stale and we pay extra for delivery.
Restaurant deliver can be worse when the driver is overloaded and your food sits at the restaurant waiting for them to pick it up, or when the driver picks up multiple orders at once (likely) and yours is delayed or is last to be dropped off.
I've definitely had this happen on occasion. There might be interesting ways to solve this problem, like showing the level of "delivery risk" associated with an order given the user's distance from the restaurant and time of day.
> It‘s generally faster than the app model where you need to wait for the driver to go to the restaurant to pick up the food
That's not how it works. I order a lot and the driver is always there to pick it up as soon as it's ready. The driver obviously gets notified about the delivery as soon as the restaurant confirms the order.
I could see how the restaurant delivery would be worse if that particular restaurant is bad at delivering. Also, the app model does seem like it should be cheaper at scale by using the deliverer’s labor more efficiently.