Uber starts same-day grocery delivery service in US

In the US it might not be as big a market due to a different lifestyle.

6. They have pretty low margins on a lot of their inventory.

I just did some back-of-the-envelope math and determined that the minimum time it could take for me to go to the grocery store and come back, from door to door, is about 40 minutes, and the maximum is about 2 hours. I took into account everything from walking to my car in the parking garage and getting out of the garage to finding a parking spot, waiting in line to check out, hitting traffic lights on the way there and back, etc.

I think this is why I don't cook as much as I'd like to. I can either spend 40 minutes getting groceries and then needing to cook a meal and then eating it, or I can spend 40 minutes picking up, say, Chipotle, which includes eating the food.

Before I started getting my groceries delivered, I would find myself regularly ordering delivery food (also online via foodler) and spending way more money because I was too tired/lazy to run to the grocery store after work.

Sure, you can use a large supermarket, but I'd really love to go online, see my local store(s), and click to purchase a handful of items and have them delivered within the hour as conveniently as a pizza.

It's one of the many ideas I'd build if I weren't already building stuff ;)

I could easily save the £5 delivery charge or whatever by having access to them thousands of items.

I don't know that this is true. I consider going to the grocery store a creative endevor about half the time I go there. Sure I start with a list and some ideas. But then there are sales on items I like. Or they have some particularly good looking produce that I want to feature in a meal. Or they are out of some other key ingredient for a recipe.... not to mention I'll be updating the menu in my head as a wander the aisles - Seeing and being in the food is extremely distracting since I really like to cook.

Sometimes I just want to go in, grab some basic supplies and leave. I will be exploring delivery options for this case, because it will be harder for a 10 minute trip to the store to turn into a 40 minute menu binge for my impromptu dinner party tomorrow.

To your specific points:

1. Locality is a huge competitive disadvantage. Real-estate, staffing, infrastructure, etc.

2. More-efficiently managing a wide selection and optimizing for product that frequently turns over is a key advantage of warehouse/delivery operations.

3. It makes no sense to compare an interruption of your day with an asynchronous order/delivery method. No-one uses Netflix the way they used Blockbuster. No-one misses Blockbuster.

4. Personally picking fruit directly off plants was more accurate than personally picking fruit in a grocery store. It's also something most people don't miss as a regular chore. Warehouse/delivery has a massive advantage here as well, as they'll be able to better manage inventory flow and don't have to suffer the same exposure to elements/contaminants as a grocery store.

They can even offer "ugly produce" at a discount, bringing the same implicit assumption of responsibility for odds and ends and generating sufficient extra revenue from the particular buyers, to effectively serve them and any returns.

There's no "niche" for this. The niche will be serving the people who enjoy farmer's markets and those who dropped a dozen eggs and will pay a premium for their replacement.

I see it as very different from something like, say, Instacart not because of the core idea of grocery delivery, but because Webvan made some mistakes that seem made-up when read today. Recall they had their IPO just 60 days after their first webstore which in turn was something like one month after their beta launch.

So they went from beta with 1000 users to IPO in like 90 days. That's, what, the time Instacart spent just in Y Combinator growing and learning?

Similar to AWS, really. Originally the AWS idea was that Amazon needed those extra servers and a way to manage them anyway, might as well do something valuable with the excess capacity.

There isn't much money in grocery delivery services at scale.

But, that's what beta tests are for. Here's hoping the next iteration gets it right.

San Francisco, by contrast, is almost too perfect. It would no doubt be successful there, but it doesn't give them as good of a feel for how this service would fare outside of Silicon Valley where these types of ad-hoc delivery services are far more common.

I get the sense that they're trying to gather data without drawing a lot of attention to their experiment.