Home delivery versus going to stores: how does fuel usage for transport compare? Any studies available?

Question

We see more and more trucks riding through our streets delivering individual items. This hardly looks sustainable.

I wonder how the transport footprint compares with the 'old' delivery model of goods being shipped to stores and bought by customers going there.
Please ignore stuff like packaging (the question is already complicated enough).

Assume the following (current) situation:

• We start at vendor warehouses across the country. Goods arrive there from the manufacturers. This first part does not show large differences. We ignore direct manufacturer to customer delivery, this happens much less frequently.
• From the warehouses goods go either to stores (store scenario) or are delivered to individual households (doorstep scenario). Pick up locations are still uncommon.
• Vendors use multiple delivery services; these do not combine their streams
• Transport is done by (delivery) truck
• Culture is European/American. Societies are wealthy enough that most people have a car and there are good transport systems.

In the store scenario we (can) go to the stores using different modes of transport. We often combine several shops in one trip. This is a major difference with the doorstep scenario in cities/towns - where most people live nowadays.
Note that US cities are much more car oriented than Europe, so the answers may be limited to one of the two.

Many factors are involved of course: population density and city design translate to distances. This will determine what transport people use.

I'm not so much interested in opinions as I am in data: have studies been done on this? What are the results?

Notes:

1. If I have forgotten major assumptions or factors please comment and I will edit them into the question
2. There is the very much related question How do I calculate if it is more sustainable to have my groceries delivered? but that also includes packaging and assumes individual car transport only. The current two answers there address only the bags and refute the car assumption.

Answer 1

Unfortunately there are not many studies available on this subject and, as you wrote, the answer depends a lot on the country and on the city.

This article on the Guardian analyses the study Impacts of home shopping on vehicle operations and greenhouse gas emissions: multi-year regional study. This study investigates the effects of home shopping on vehicle operations and greenhouse gas emissions in Newark.

Simulation results showed that home shopping will put additional burden on Newark transportation network, as identified through four measures of effectiveness (MOEs) which were travel time, delay, average speed and greenhouse gas (GHG) emissions.

According to the researchers one of the reason is that people use the time saved by doing online shopping in doing other activities that involve using a car.

“We found that the total number of vehicles miles travelled hasn’t decreased at all with the growth of online shopping,” says study leader Arde Faghri, professor of civil and environmental engineering and director of the Delaware Center for Transportation. “This suggests that people are using the time they save by shopping on the internet to do other things like eating out at restaurants, going to the movies, or visiting friends.”

A more general study can be found in Home Delivery and the Impacts on Urban Freight Transport: A Review but the results are not really conclusive

Some of the changes result in increased pressure for road traffic networks in sensitive areas (for example residential areas). At the same time the changes also provide opportunities for the use of vehicles powered by alternative fuels thereby supporting certain sustainability strategies. However, the changes are complex and patterns are not the same from one country to another thus the impacts on city logistics are also varied.

Answer 2

Much depends on the density of deliveries.

Consider:

Suppose that the average distance between customers was 1 km, and the warehouse is 10 km away. A simple approximation says 9 km driving between customers, and 2 10 km trips from and back to the warehouse. Net travel 30 km.

If each customer has to travel to the warehouse, you have 10 * (10+10) = 200 km.

But as you pointed out customers are likely to combine errands. But I seriously doubt that many people will combine 10 or more errands.

Alas, it's not that simple:

If a customer is out for a non-deliverable, say, taking Susan to soccer practice, then it may still make sense for the customer to pick up at warehouse.

For the vendor, it may make sense to consolidate deliveries by time. Say NW Edmonton only gets deliveries on Tuesday, then instead of 10 customers you have 50. The average distance drops by a factor of 5, at the price of people needing to wait longer.

It would also make sense for people to have drop boxes. E.g. Elon Musk sets up a new company that provides 1 cubic meter boxes for rent on an hourly to monthly basis. Acme Widget drops off my new widget, not to my house but to Hold'em Warehouse. Employess put it in my box for later pickup.

A day later my order from Bilbo's Bookstore arrives. It's added to my box.

Three hours later Superstore foods drops my groceries off. They go in my box, and an alert goes to me: Milk delivered to your box. Perishable.

At this point either I go by an empty my box, or I phone up EM Transport, and arrange for it to be delivered.

By setting up distribution networks like this you reduce the delivery costs, but increase the handling costs.

Not sure where the optimum balance point is.

In Canada we are going to "Super Boxes" Instead of delivery to your door, there is a superbox usually at the point where your road connects to distributor road. A set will server about 30-60 houses. Boxes are available in various sizes.

When we get a parcel that is too big for our box, the postman leaves a card for you to pick up at the local postoffice.

We are on Amazon Prime. Most of the in country shipping is free, and most arrives in 3 business days. (They say 2, but don't count the day you place the order) But Amazon is big enough that whole semi-trailers/747's go to major points.

To get decent rates from the post office here you have to have volume. And your packages have to be sorted by postal code. We have a local tree company, tree time, that during the shipping season in spring is sendout out 2 27 foot vans of parcels a day. But they have to pallet them by postal code