What is the most efficient way to organise grocery retailing?
Image sources: ICA & Unsplash
There has been long and sometimes heated debate around what is the most efficient way to organise grocery retailing.
Online start ups argue that building and operating a couple of warehouses per metropolitan region is much more efficient than building big amount of stores.
The store based retailers see that the self service remains an undisputed model in efficiency as the customers do the most expensive parts of the work: picking the products and delivering them home. Retailers just need to stock the shelves. This model has been optimised to perfection over the last decades.
Which one is more efficient?
We had an opportunity to get data from both an online pure play retailer as well as from a hypermarket operator. We set out to find, which one is more efficient operate
pure play home delivery online retailing OR
traditional self service hypermarket (without the online business)?
One has to note that online grocery has remained (for so long) far less profitable to the traditional model.
One key reason for lower profits online has been the fact that online is still operating on small volumes.
These small volumes apply to sourcing as well as logistics.
In sourcing it will still take time for the online players to reach such levels of sourcing that they can be competitive against the incumbents. For logistics, the volume gains can be reached earlier for picking and delivery.
After all grocery retailing is a high volume business where the economies of scale have a big influence in driving the profitability. Very few hypermarkets could operate profitably with only 25% capacity.
Comparison
When comparing the profitability of the two models one has to differentiate building and runing the businesses. We have compared the efficiency in both phases separately.
Before we dive into the details some basic information is needed
to compare the two businesses we have used 300 million € as the basis for calculating the costs
this would be roughly the scale for an online grocery warehouse to operate efficiently
same amount of sales would require on average five to six normal hypermarkets
to compare this to Finnish or Swedish hypermarket chains
K-Citymarkets sell on average 31,3 M€
Prisma 47,2 M€
ICA Maxi 47,8 M€
Thus it would be fair to say that it takes approximately six hypermarkets to generate 300 M€ in revenue.
Set up costs/building costs
We estimate that the building costs of a new hypermarket would be around 35 % smaller than the building costs for one warehouse. Naturally the building costs can vary significantly depending on whether the store is a new development or in an existing building.
On the other hand, setting up six hypermarkets is (significantly) more expensive than building one warehouse.
It costs almost four times more to build the equivalent amount of hypermarkets than one warehouse.
Setting costs for the warehouse can vary a lot. Ocado warehouses can be used as a measure for the upper end of the price spectrum. The Ocado warehouses are probably the most automated. ICA invested 110 M€ for the Ocado warehouse. Even that would be cheaper than building the six hypermarkets. The biggest and most expensive warehouses also can handle more revenue than 300 M€, which is used as a basis for this calculation.
Another major difference in the building phase were the cold machines.
Cold machine investments were roughly ten times more expensive for the hypermarkets than for the warehouse.
In hindsight this is understandable as the warehouse can be equipped with much more bare bones solutions than the stores.
After all warehouses are designed only for the picking efficiency whereas the stores are designed to create a more or less pleasant experience for the customers who come to the store. This includes elements that add costs quite significantly, such as service counters, lighting...
In terms of equipment investments, the warehouse was more expensive in total as there is automation and other machinery that are built into a modern online grocery warehouse. Another big difference is the investment in delivery trucks for the warehouse. That creates a several million euro extra cost for the warehouses.
One area, which this calculation does not even take into account is all the work in finding a good slot and getting permissions to build on it.
In many cases this is not done by the retailer, but by the property developer. However, this takes a lot of time for six hypermarkets, because location is the most critical individual element in the success of the store. For the warehouse this development process needs to be done only once.
Costs of running the operation
How about then the daily costs of running these operations?
Surely self service retailing concept where customers drive to the store and do the picking of the products is cheaper than the online version.
With low volumes (around 50 M€) the warehouse costs a bit more to operate than one hypermarket. The difference on low volumes is -9% compared to one hypermarket. For the warehouse costs amount to 18% of revenues, whereas for the hypermarket costs make up 14% of revenue.
However, the balance changes as the warehouse starts to operate at higher volumes.
With full volumes, the warehouse has operational costs of only 8% of revenues. For six hypermarkets it costs 70% more expensive to operate the hypermarkets.
The biggest differences in the running costs arise from personnel costs and rents for the hypermarkets.
Personnel costs can be up to 100 % bigger for the hypermarkets, because of the labor intensive nature of the many processes in the stores. Putting products on display, serving customers on the service counters and cashiers are probably the biggest sources of labor in the stores. All of this work is neither totally removed or done significantly more efficiently in the warehouse.
The rents are another major differentiator in the running costs. Hypermarkets are always located in a more convenient location for the customer to access the store. In Finland many hypermarkets are even located in shopping centers. Whether comparing to a free standing out of town hypermarket or a shopping center hypermarket, warehouses are located in significantly cheaper land thus making the rental cost notably smaller.
Summary
Whether it is surprising or not, online grocery retailing from the warehouses seems to be more efficient way to organise grocery retailing
Online grocery has been seen as a far more unprofitable medium for selling groceries.
However, grocery retailing is a volume business where increasing the top line makes operations much more efficient. That is why the grocery retailers tend to be very big companies.
Online grocers are still in many ways in their infancies, young companies still waiting for the economies of scale to kick in. Hypermarkets would not be efficient if they were to operate at 25% capacity.
The good news for the warehouse based online grocers is that the warehouse model should scale up quite efficiently. Costs of running a warehouse at low capacity (around 50 M€) are a bit less than 20% of revenue. For a hypermarket running at an average capacity (around 50 M€) costs make up 15% of revenue
As the warehouse scales up to higher volumes, the share of revenues going to costs decreases. Online is also easier and more efficient way to scale up geographically.
Kroger's online only expansion into Florida and Texas are interesting experiments to follow in this domain.
How about sustainability?
Another rather overlooked aspect of online grocery is the environmental impact of one big warehouse compared to many smaller hypermarkets.
Online has an advantage environmentally in terms of reducing the amount of driving. Instead of all customers needing to drive to a grocery store, one delivery truck can deliver 20-30 customers in one, much more efficient, route.
This saves a lot of driving and thus is better for the environment. One could argue that online grocery is more efficient in terms fo driving, when the orders are picked from the stores. This could be true, if the order volumes were high enough to make the routes efficient. Then the routes would start closer to the customer.
However, then one should also consider the energy usage of the stores. It is likely that the warehouse uses a lot less energy than the six hypermarkets combined.
More efficient, but not necessarily better
More efficient is not the same as better. It boils down to customer preferences. And we will have both options for customers to choose, neither is disappearing anytime soon.