How do supermarkets make money?

As I walked home from Whole Foods this week I wondered just how different John Mackey’s stores are from the rest of the food retailing industry. On the surface they seem very different – Whole Foods is richly appointed relative to the other supermarkets I frequent. How can they afford to do this?

To answer this question, I decided to take a look at a few financial operating metrics for Whole Foods and a set of publicly traded competitors. Since I’m interested in operating difference, I chose to start by examining each firm’s return on assets (ROA).1 Data is readily available for:

  • CostCo – Membership retailer focused on low costs. Groceries are a significant portion of volume.
  • The Fresh Market – Supermarket with similar appearance to Whole Foods. Footprint primarily on the east coast.
  • Delhaize – Global supermarket operator with several grocery brands in the eastern US. Brands include Food Lion, Hannaford and Sweetbay.
  • Publix – Employee owned supermarket chain in the southeastern US
  • Safeway – National operator of large supermarkets
  • Kroger – National operator of large supermarkets under multiple brands

Unfortunately, it is difficult to disaggregate Wal-Mart’s grocery performance from the rest of their business. Given that Wal-Mart is the top food retailer in the US by sales volume, excluding it will admittedly leave out one of the most important players. On we must go…

Supermarket ROAs

Source: Company filings

As you can see, Whole Foods (in green) is among the best performers in this set of supermarket competitors. Performance slipped during 2006-2008 as Whole Foods struggled to raise the performance of over 100 recently acquired Wild Oats stores. While this begins to tell the story, we can dig a bit deeper to gain a better understanding of what has driven the difference in performance between these companies. Specifically, the net income margin and asset turnover can shed light on how profitable each supermarket is, and how well each utilizes its assets.2

Supermarket net income to sales

Source: Company filings

Supermarket asset turnover

Source: Company filings

Competition among supermarkets is fierce – most mass-market players have net income margins between 1 and 2% of sales. Only Whole Foods and Publix have consistently achieved greater margins. It is interesting to note that the Wild Oats acquisition more than halved Whole Foods net income margin from 3.5% to 1.5%.

Whole Foods performs near the bottom of its peers in asset turnover, however. CostCo achieves relatively high turnover with its concrete floors, metal shelving and huge pack-sizes. Kroger has improved significantly over this time frame, although on the surface the drivers of this improvement are unclear. Either Kroger has become much more efficient, or they have allowed their assets to depreciate with low reinvestment or maintenance expenditure.

Margin performance can be further disaggregated3 into:

Supermarket COGS to sales

Source: Company filings

Supermarket SGA to sales

Source: Company filings

Here we start to see the true differences between Whole Foods and other supermarkets. Whole Foods has the lowest cost of goods sold (COGS) among its competitors, which is synonymous with a greater markup above what Whole Foods pays for its products.In addition, Whole Foods has the greatest selling, general and administrative (SG&A) costs as a percentage of sales.4  This money is likely spent on additional staff in each store to provide a high level of service. Alternatively, CostCo has by far the greatest COGS and therefore the lowest markups. In addition, CostCo spends very little on SG&A. These financial metrics align perfectly with each company’s strategy. Whole Foods provides high quality products at a higher price point delivered with a high level of service, while CostCo focuses solely on delivering products at the lowest possible price. In fact, CostCo derives most if not all of its profit from membership fees. Products are priced to simply cover operating expenses.

Asset turnover can similarly be disaggregated:

Supermarket inventory turnover

Source: Company filings

Supermarket noninventory asset turnover

Source: Company filings

Again, the finer level of detail uncovers differences between Whole Foods and the competition. Whole Foods (and The Fresh Market) turn their inventory significantly faster than other supermarkets. This performance is likely driven by a greater proportion of perishable goods (produce, meat, seafood, deli items) which drives more frequent shopping trips. Additionally, Whole Foods’ low non-inventory asset turnover (COGS / non-inventory assets)5 implies that Whole Foods provides a more luxurious (more expensive) shopping experience. CostCo again provides a useful contrast. Inventory turnover is middle-of-the-pack, while non-inventory asset turnover is significantly higher than others. Whole Foods provides a very luxurious shopping experience for the grocery customer, offset through greater inventory turns. CostCo focuses on a no-frills shopping experience, and therefore has very high non-inventory asset turns.

Bringing it all together allows us to examine the impact each operating driver has on supermarket performance:

Supermarket ROA tree

Source: Company filings

Overall, Whole Foods has constructed a unique operating model relative to other supermarkets. They have chosen to focus on high-margin, faster turning perishable goods, which provides the money and operating flexibility to provide both better service and a more luxurious shopping experience. It will be interesting to see whether Whole Foods can fully raise the performance of the Wild Oats stores to Whole Foods standards, as well as whether The Fresh Market can provide a legitimate competitive threat.

Bonus: Equity performance of a few supermarkets since 2002. After being hit hard by the recession and the Wild Oats acquisition in 2007-2008, Whole Foods has made an impressive recovery. One could imply that the market is reflecting a consumer trend toward either high-end (Whole Foods) or low-cost (CostCo) food retailers.


  1. I should note that return on invested capital (ROIC) is typically considered a better metric than ROA for analyzing a potential investment in a company. However, in company’s with similar capital structures and working capital needs, ROA and ROIC are similar. In addition, ROA can be easily disaggregated into granular components to gain a better view on the drivers of performance. [return to reading ↩]
  2. Note that (net income margin) x (asset turnover) = return on assets, or (net income / sales) x (sales / total assets) = (net income / total assets) [return to reading ↩]
  3. 1 – COGS/Sales – SG&A/Sales = Net income/Sales [return to reading ↩]
  4. For the purposes of this chart, SG&A was calculated as Sales – COGS – EBIT [return to reading ↩]
  5. Note that non-inventory asset turnover is a non-traditional metric, but it provides a useful proxy for the quality of the infrastructure that supports the shopping experience. In this analysis, non-inventory asset turnover is calculated as (COGS) / (total assets – inventory) [return to reading ↩]

Making sense of sustainability in a world of complex systems

Some days the more I think about food systems, the more confused I become. There are so many components, each with complex and ever-changing connections to other pieces of the puzzle. Each time my inclination is to sit down with a pen, paper, and the determination to finally put everything down in neat little boxes. Unfortunately, it hasn’t been so easy. Change in the system will come from not only understanding where food, money and information flows, but also where influence arises. I have yet to stumble upon a comprehensive attempt to catalog the interdependence of the system (if you have one, please send it along!). But I must try, if simply to provide a framework through which to collect my thoughts.

To start, I wrote down all the major components of the U.S. food system.

Food system jumble


My first step in organization was to separate those pieces traditionally associated with business from the others. For these purposes I used a broad definition of business to include any individual or group of individuals creating, buying or selling food. Therefore the business category includes everyone from a local farmer at a farmers market to the largest agribusiness corporations.

Business vs other

These components make up the traditional food supply chain, and can be consolidated for simplicity’s sake. Farmers and ranchers represent those acting as producers. Packagers, distributors, restaurateurs, and retailers comprise the primary supply chain functions. (Note that even with the high-level categories I began with there is still a need for refinement to create an easily understood graphic. Unfortunately, this will continue. I’ll revisit this issue later.)

Business vs other refined

Now that the business system has been identified (light grey box below), I can begin adding in the other pieces. First up are the last stop in the supply chain, the consumers.

Traditional supply chain

Next, I can add the members who have influence on the supply chain but are not directly involved. Government creates institutions and a regulatory environment in which businesses and consumers live. Non-profits provide valuable research to the system, and may often exert strong influence. Business, consumers, government and non-profits collectively form a complex community of people and organizations (light green box below).

Business consumers and community

Finally, the food system community is embedded in the environment (green box below). The system both relies on and has an impact on natural resources and eco-systems found around the world.

Food system

Each of these boxes represents an actor in the system, and a web of interdependence connects each actor to all others. This highly complex, dynamic interdependence is where I continually run into a dead end. Therefore, rather than attempt to visualize each and every link in the system it would be useful to develop a simple reminder that all this interdependence exists. To do this, I’ll borrow the pentad framework from Carol Sanford’s The Responsible Business. Although it uses an organizational focus to view the system, it can be loosely applied to show the connections among all actors.


I find this framework to be particularly useful as a high-level reminder that any change to one component of the system will affect all others. Focusing all attention on supply chain partners, for example, leaves four other key components from the plan for change.

Combining business-related components into one allows even further refinement. I end up back at the broad definition of sustainability seen earlier (in this post):

sustainability venn diagram

At the end of all this I take away two things:

  • The food system is very complex, and the strong interdependent relationships imply that change cannot simply be focused on one component.
  • Visualizing, diagramming, or otherwise mapping the food system in ever-increasing detail will more than likely provide diminishing return. A more efficient approach would take individual deep analysis and place it in the context of the larger system.

What are the environmental consequences of growing the food supply to feed the world in 2050?

The most recent UN estimates show global population stabilizing at 9-10B sometime after 2050. What will it take to feed these additional 2-3B people?

While forecasting anything 40 years in the future is a treacherous task, estimates typically place the required expansion of the food at double current production levels in 2050.1 I often find it easier to visualize some of these data to get a better sense of the magnitude:

Data from UN sources. FAO statistics database and population database

It is interesting to note that during the past 50 years, otherwise known as the green revolution, global food supply has grown faster than global population. Thomas Malthus must be quite happy that his 1798 prediction that the food supply was limited to linear growth has proven incorrect, at least thus far. Malthus failed to predict that increases in yield per cultivated area would increase simultaneously with total area under cultivation. In other words, increases in the food supply are driven by:

The first term, yield per cultivated area, has driven a majority of the expansion of the global food supply in the past 50 years. Future improvements will also likely lean on this lever.

Understanding of the potential environmental impacts of expanded food production  requires bit deeper dive into the factors that will drive the increase. Luckily others have already completed great work on this topic. Tilman et. al. lay out the drivers of the future impact in their 2000 paper Forecasting Agriculturally Driven Global Environmental Change.

Yield per cultivated area is a function of:

  • Fertilizer application
  • Irrigation
  • Pesticide application
  • Yield potential of crops2

Cultivated area is a straightforward function of area under crop or pasture production.

The use of these inputs for food production in their current form also have negative environmental impacts. These impacts include:

  • Irrigation drives salt and nutrient loading to downstream aquatic ecosystems, salinization of soils, and pressure on limited water availability
  • Fertilizers drive eutrophication3 of estuaries and coastal seas, loss of biodiversity, groundwater pollution, increases in smog and ozone depletion and acidification of soils
  • Pesticide drives general environmental degradation and human health effects
  • Conversion to cropland drives global habitat destruction and loss of native ecosystems

Tilman et. al. conclude that given the current trajectory of food production, the environmental impact in 2050 will be quite large.

  • Irrigation – 2.0 fold increase in area under irrigation, placing severe stress on global water tables
  • Fertilizers – 2.4 to 2.7 fold increase in nitrogen- and phosphorus-driven eutrophication of terrestrial, freshwater, and near-shore marine ecosystems
  • Pesticide – 2 to 3 fold increase in pesticide use, leading to general environmental degradation & human health concerns
  • Cropland conversion – 1 billion hectares of natural ecosystems converted to agriculture by 2050 (about the area of the U.S.)

There is clearly a need to find solutions to ensure adequate food supply growth while simultaneously using less fertilizer, water, pesticide and land per unit of output. This is a massive but important challenge, and one which will provide ample opportunity for entrepreneurial innovation over the coming 40 years.

As a quick end note, it is interesting to examine the predictions of Tilman et. al. 10 years in to their 50 year predictions (the report was published in 2000). Using the same UN statistics, the annual use of key inputs can be examined and compared to the estimates of the study. In the figures below (please click on the images for a larger version), note that irrigated land and both types of fertilizer are closely tracking the predictions. Both cropland and pasture land, however, are falling well below predictions. I will not speculate here to the causes of this phenomenon, but it does give hope that we’ve already made progress to grow the food supply with fewer inputs.

Irrigated landNitrogenous fertilizerPhosphate fertilizer


Crop landPasture land

  1. These estimates assume an increasing shift to more protein-dense, calorie-rich Western style diets. For example, this UN report on Global Agriculture Towards 2050 reports an expected availability of 3,050 calories per capita per day. [return to reading ↩]
  2. Yield potential is the theoretical yield under ideal conditions. See this definition for more information [return to reading ↩]
  3. I thought it was a made up word as well. See Wikipedia for more information [return to reading ↩]

What does a sustainable food system look like?

All this talk about food sustainability – and I rarely know what we’re aiming for. I wonder not just what food sustainability is, but what does a sustainable food system consist of?

Where to start with such an expansive question? With the simple things. Food (as defined by google) is: any nutritious substance that people or animals eat or drink, or that plants absorb, in order to maintain life and growth. One thing that I love most is the massive variety. At the moment, Wikipedia lists 113 pages each with a different list of foods. This variety, however, expands the scope of food sustainability to a staggering breadth of items.

Sustainable is a bit more difficult, but still straightforward (again, by google): able to be maintained at a certain rate or level; or (esp. of development, exploitation, or agriculture) conserving an ecological balance by avoiding depletion of natural resources. Additionally, a system is considered sustainable if the  environmental, social and economic demands are present and balanced (from Wikipedia).

Given these definitions, what  is sustainable food? Most simply, my working definition of food sustainability is food that comes from a system where environmental, social and economic needs are balanced and can be fulfilled over a usefully long future period of time.

Perfect. But how will we know when we’ve reached this ideal? And how long of a period is enough time over which to achieve sustainability? These are the questions that I’ve struggled with and why I started this blog.

I’ll continue to shape and refine this definition in the coming weeks and months, but in the meantime here are a few items that have been influencing my perspective:

USDA legal definition of sustainable agriculture

  • Satisfy human food and fiber needs.
  • Enhance environmental quality and the natural resource base upon which the agriculture economy depends.
  • Make the most efficient use of nonrenewable resources and on-farm resources and integrate, where appropriate, natural biological cycles and controls.
  • Sustain the economic viability of farm operations.
  • Enhance the quality of life for farmers and society as a whole.

FAO (Food and Agriculture Organization of the United Nations) definition of conservation agriculture

  • Conservation agriculture (CA) aims to achieve sustainable and profitable agriculture and subsequently aimes at improved livelihoods of farmers through the application of the three CA principles: minimal soil disturbance, permanent soil cover and crop rotations.

UC Santa Cruz Center for Agroecology & Sustainable Food Systems

  • Agroecology is the science of applying ecological concepts and principles to the design, development, and management of agricultural systems.
  • Sustainable systems are environmentally sound, economically viable, socially responsible, nonexploitive, and serve as a foundation for future generations
  • Food systems are the people and resources involved in producing, processing, distributing and consuming food and managing waste – a food system operates within social, political, economic and environmental contexts.

A few essays on the topic:


Why does food sustainability matter?

In the past, I’ve struggled to articulate the reasons why food sustainability is important to me. Beyond a love for food and the connection with others that it can offer, I’ve collected a few facts over the years related to the food supply chain’s importance:

20-30% of human environmental impact is tied to food consumption, equal to the impact from housing and a greater impact than all of human transport1


Feeding a growing US and global population is a challenge – doing so with unsustainable solutions will only shift problems to the future when further solutions will be required

  • The Economist predicts 9 billion people will need to be fed daily by 20502
  • Agricultural productivity will likely have to outpace the rate of improvement seen in the past 50 years to meet demand in the next 50 years3


Cereal productivity improvement needs to outpace past performance to meet future demand (Tester & Langridge, "Breeding Technologies to Increase Crop Production in a Changing World")


The food sector hosts some of the poorest working conditions in the US. I admittedly have fewer facts to support this issue, but some interesting reads include:


  1. This conclusion emerges from a study of 11 studies which assessed impact based on varying combinations of GHG emissions, land use, environmental interventions, total energy requirements, or life-cycle impact assessments (LCAs). Despite varying methods, each report indicated food accounted for 20-30% of total impact (housing 20-30%, human transport 15-25%) Note that the scope of this study focused on Europe, therefore limiting its application to the US in a number of ways. Most prominently, Europeans live in smaller houses in a more geographically dense area than Americans. This could skew a larger portion of American consumption to housing and transport. However, food will likely remain a prominent component of environmental impact. See the report here: Tukker and Jansen – Environmental Impacts of Products. [return to reading ↩]
  2. The Economist – The 9 billion people question. [return to reading ↩]
  3. Tester & Langridge, “Breeding Technologies to Increase Crop Production in a Changing World”. [return to reading ↩]

First post


This is the first of what I hope to be many frequent postings to share my thoughts and research on the US food supply system. The content will inevitably jump around, but my goals for the site are the following:

  • Examine the path that food takes from soil to plate
  • Identify high potential areas ripe for sustainability improvement
  • Develop profitable business models to facilitate change in these areas

Along the way I hope to draw from the expertise of the broader food community to push this thinking forward.

I’m looking forward to the journey!