This seems like such a solvable problem. It only requires some profile setup.
Take the most commonly re-ordered items on Amazon (paper towels, toothbrushes, etc) and have the user select a preferred product, brand, and fallback for each one, as well as a maximum price. Some of that is already implemented for Subscribe & Save and Whole Foods delivery/pickup orders. They could even leverage that existing system.
For less common items, first check the user’s purchase history for something similar and announce if the price has increased more than few percentage points or is not in a normal range for the product category.
That will cover a lot of cases. For the rest, refer the user to do some research on the computer.
Nobody wants to go through building a list of common items. I would just leave the door and come back with my toilet paper, bought from the shop, 5 minutes later.
Also I don't understand that "let's order household stuff" online randomly and individually example. Sure sometimes you run out of something but I think it mostly happens to people like me who live in small places because we don't have the space to have 3 packs of 6 rolls of toilet paper. But usually people who live in small spaces are usually the ones living in high density places where small groceries shops are always available at walking distance. People who live in suburbs usually have enough storage space and do their shopping of houshold items once a month and never run out of anything.
I have most household items on subscription. It saves a ton of time and costs about the same, maybe a little less for some items. And most importantly, it means I never forget something and then have to make another trip.
For me, in spite of living in an ex-urban location, there's a Walmart a 10 minute drive away. But I can absolutely see (and know people) for whom picking up a big package of paper towels in Brooklyn is sort of a pain.
Take the most commonly re-ordered items on Amazon (paper towels, toothbrushes, etc) and have the user select a preferred product, brand, and fallback for each one, as well as a maximum price. Some of that is already implemented for Subscribe & Save and Whole Foods delivery/pickup orders. They could even leverage that existing system.
For less common items, first check the user’s purchase history for something similar and announce if the price has increased more than few percentage points or is not in a normal range for the product category.
That will cover a lot of cases. For the rest, refer the user to do some research on the computer.