[cehrnrooth]

Location density of the deliveries has got to have a massive impact on their effectiveness. I'm highly skeptical of their ability gather groceries and deliver to 5 separate locations within 1 hour. Even in a dense area where it's 5 minutes between each location (plus parking, plus entry / exit and delivery) that's 25 minutes minimum.

The longer the time, the lower the $/Hr. The further away the locations are, the higher the transportation costs are for the driver, further cutting into their actual take home pay (not including, miles, depreciation, insurance, etc...).

Instead of making deliveries after the order (and trying to bundle orders and deliveries), I would think a dispatch model where one set of workers go out in the morning and make the orders and deliver to a central location. Then have a second shift make deliveries, but that would make it difficult to maintain the independent contract status of their workers.

[Dwolb]

Agrees on the location density point - the route for the driver has got to be near perfect for their model to work.

The dispatch system is an interesting thought. I wonder if it fully aligns with their delivery time promise.

I think the dispatch system points to another 'lever' however - certain recurring orders can be lazily filled. i.e. when the shopper has no rush orders he/she can fill her free time filling orders which are upcoming, but not immediate