Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

RESEARCH

Research Paper

Food Marketing Practices of Major Online


Grocery Retailers in the United States,
2019-2020
Alyssa J. Moran, ScD, MPH, RD; Gabby Headrick, MSPH, RD; Crystal Perez, MPH; Alexandra Greatsinger, MSc; Lindsey Smith Taillie, PhD;
Laura Zatz, ScD, MPH; Sara N. Bleich, PhD; Eric B. Rimm, ScD; Neha Khandpur, ScD, MSc

ARTICLE INFORMATION ABSTRACT


Article history: Background Food marketing influences consumers’ preferences for and selection of
Submitted 1 August 2021 marketed products. Although a substantial body of research has described food-
Accepted 7 April 2022 marketing practices in brick-and-mortar stores, no research has examined food market-
ing in online grocery retail despite its growing importance as a source of food-at-home
purchases.
Keywords: Objective To develop and apply a coding instrument to describe food marketing and
Marketing
the nutritional quality of marketed products in online grocery stores.
Online retail
Supermarket Design Quantitative content analysis and review of product Nutrition Facts labels and
Food environment ingredients lists to calculate nutrient density and level of processing using the NOVA
Food industry classification system.
Supplementary materials: Participants/setting Foods and beverages (n ¼ 3,473) marketed in the top revenue-
Table 5 and Figure 2 are available at www.jand.org generating online grocery retailers and those participating in the US Department of
Agriculture Supplemental Nutrition Assistance Program Online Purchasing Pilot (n ¼ 21)
in 2019-2020.
2212-2672/Copyright ª 2022 by the Academy of
Nutrition and Dietetics. Main outcome measures Use of marketing mix strategies (ie, product, placement,
https://doi.org/10.1016/j.jand.2022.04.003 promotion, and pricing) across retailers and nutritional quality of marketed products.
Products were considered of poor nutritional quality in the case that they were ultra-
processed (NOVA category 4) and excessive in sodium, saturated fat, free sugars, and/or
other sweeteners. Products were also classified into 13 mutually exclusive food groups.
Statistical tests performed The proportion of retailers using each marketing strategy,
proportion of products of poor nutritional quality, and proportion of products in each
food group were calculated.
Results Retailers commonly used product recommendations, search result ordering,
branded website content, user-generated content, and social media engagement to
market products online. Candy, sweets, and snacks made up the largest percentage of
marketed products (17.3%), followed by fruit, vegetables, and legumes (16.7%). Most
(62%) marketed products were of poor nutritional quality. Staple food categories such as
fruits, vegetables, and grains were frequently marketed, particularly through price re-
ductions and product recommendations.
Conclusions Online grocery retailers use a variety of customizable food marketing
strategies on their websites. Although most marketed products are of poor nutritional
quality, there is potential for marketing of staple food categories online that is not
feasible in a brick-and-mortar store.
J Acad Nutr Diet. 2022;122(12):2295-2310.

S
INCE THE START OF THE CORONAVIRUS DISEASE younger shoppers, stores are expanding to populations
2019 (COVID-19) pandemic, participation in online historically underserved by access to healthy food.2 For
grocery shopping has risen dramatically. In 2019, e- example, Amazon offers discounted Prime memberships to
commerce accounted for only 3.4% of the US gro- Supplemental Nutrition Assistance Program (SNAP) partici-
cery market; by 2020, that number had grown to 10.2%.1 pants,3 and the US Department of Agriculture (USDA)
Industry projections estimate that 90% of current online recently expanded the Online Purchasing Pilot, which al-
grocery shoppers will continue to shop online after the lows online payment with SNAP, from five states in March
pandemic, with online sales accounting for more than 20% 2020 to 47 states and Washington, DC, by March 2021.4
of the grocery market by 2025. Although online grocery Despite the rapid growth in online purchasing, very little is
shopping has been most popular among higher-income, known about food marketing in US online grocery stores, or

ª 2022 by the Academy of Nutrition and Dietetics. JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2295
about the nutritional quality of marketed products. In brick-
and-mortar stores, food manufacturers and distributors pay RESEARCH SNAPSHOT
incentives (“trade fees”), amounting to more than $50 billion
Research Question: What are the marketing strategies used
each year, to market predominantly unhealthy products
by the top revenue-generating online grocery retailers in
through in-store advertising, price discounts, and displays.5-7
This in-store marketing is effective, and has been shown to
the United States to promote food and beverages to
increase preference for and selection of advertised foods.6,7 In consumers, and what is the nutritional quality of these
contrast, online platforms lend themselves to food marketing products?
that is highly flexible and customizable at the individual Key Findings: Online grocery retailers commonly combined
level.8 In this setting, consumer data can be leveraged to marketing strategies such as product recommendations,
create personalized price discounts, product recommenda- paid product placements, and branded website content to
tions, and web content.8-11 This targeted online marketing
promote products to US consumers. Nearly two-thirds (62%)
may increase the effectiveness of traditional strategies to
of marketed foods and beverages were ultraprocessed
influence consumer decisions, resulting in increased
(NOVA 4) and excessive in nutrients to limit; although fruits,
spending and, depending on the nutrition profile of marketed
products, increased selection of unhealthy food. Online food
vegetables, and grains were frequently marketed online
retail marketing may also encourage unhealthy purchases in through price reductions and product recommendations.
new ways. For example, e-commerce sites have been found to
use tactics such as sneaking (adding items to the cart without
users’ consent), scarcity (eg, low stock messages), and social marketing strategies.18 Codebook items were selected in
norms (eg, “27 people are looking at this product right now”) alignment with Khandpur and colleagues’s8 conceptual
to manipulate shoppers into making purchases they other- framework describing the influence of the online food retail
wise would not make.12 Online retailers have also used environment on consumer behavior and included four do-
strategies such as placing products in the first row of search mains: retailer policies and practices, food marketing, web-
results, suggesting product substitutions, and offering site customization, and equity and transparency. The
personalized recommendations at checkout (so-called cart Khandpur and colleagues8 framework describes constructs
intercepts) to steer consumers toward certain products or and examples of website characteristics within each domain
brands.8,9,11,13 based on an extensive review of the academic and industry
Prior work describes a framework for how the online food literature, key informant interviews, a mock shopping exer-
retail environment may influence food choices.8 But, there cise, and expert discussion. For example, retailer policies and
has been little research documenting the nature or extent of practices includes constructs such as site access (ability to
marketing strategies or the types or nutritional quality of sign-up and cancel memberships online, incentives for free
marketed products in major US online grocery retailers.9,11,14 trials, and mobile-friendly websites), privacy and data
This project had two aims: to develop an instrument to code sharing (use of personal data), inventory management
website features and marketing practices of online grocery (product availability and price point), and grocery collection
retailers; and to apply the coding instrument to describe the and payment (delivery availability and fees and accepted
volume and content of food and beverage marketing, as well payment methods). Food marketing includes constructs
as the nutritional quality of marketed products, in 21 major organized into four subdomains based on the classic mar-
US online food retailers and retailers authorized for the USDA keting mix: product, price, placement, and promotion. The
SNAP Online Purchasing Pilot. product subdomain includes product mix (ie, assortment of
products that can be viewed on the site); the price sub-
MATERIALS AND METHODS domain includes discounts, rewards, and time-limited details
Selection of Online Grocery Retailers (eg, a weekly flyer); the placement subdomain includes
Retailers were purposively sampled to reflect the range of production recommendations, the order in which search re-
web features and marketing practices used by full-service sults are presented on the site, product advertisements, and
supermarkets that have an e-commerce platform and were other branded site content; and the promotion subdomain
on Euromonitor’s list of the top-20 revenue-generating re- includes user feedback (ratings and reviews), social media
tailers in 2018,15 or were approved for the Online Purchasing interaction, and point-of-purchase information (eg, food la-
Pilot as of September 2019.16 The Online Purchasing Pilot was beling). The website customization domain includes con-
originally authorized under the 2014 Farm Bill and included structs such as product information display (sort and filter
eight retailers at the time of this study.4 Twenty-one chain features that allow customers to view products according to
retailers were included. Because a specific store location must certain attributes), site navigation (eg, features that allow
be selected to place an online order, the store closest to the customers to customize the look and feel of the site), and
address of the retailer’s headquarters that also had a food shopping tools (eg, shopping lists). Equity and transparency is
pantry accepting donations within its delivery area was a cross-cutting domain that refers to the differential influence
selected for inclusion (Table 1). The food pantry’s address was of web features on historically marginalized populations and
used to complete the shopping task and provided an anon- information disclosure by the retailer. For example, reduced
ymous location for grocery delivery. membership fees for SNAP participants is a feature falling
within the equity domain, whereas disclosure of product
Measure Development: The Online Grocery Retail sponsorship is a feature falling within the domain of trans-
Environment Coding Instrument parency. The two lead researchers reviewed this framework
Step 1: Identification of website features. A structured and several popular online food retail websites to develop a
coding instrument was developed to assess the online gro- list of product marketing features to include in the coding
cery retail environment, including website features and instrument. Each marketing domain included dichotomous

2296 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
Table 1. Characteristics of stores included in a content analysis of 21 US grocery retail websites, 2019-2020
December 2022 Volume 122 Number 12

Demographics of Store Zip codeb


Median White Black Hispanic
E-commerce Specific store location Median household High school Povertyd population population population
Retailer name platforma (city, state, Zip code) c
age (y) income ($) or higher (%) (%) (%) (%) (%)

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ%ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Albertsonse Instacart Boise, ID, 83705 35.1 46,045 92.3 19.4 81.8 1.6 11.3
e
Aldi Instacart Aurora, IL, 60506 35.0 61,231 79.8 14.5 62.8 12.7 39.7
Amazon Fresh/ N/Ag Seattle, WA, 98126 38.0 80,142 92.1 14.4 66.5 10.9 7.0
Primef
Amazon Whole N/A Austin, TX, 78744 29.4 49,633 71.8 22.8 67.8 7.9 75.3
Foodse
Dash’s Marketf Rosie Buffalo, NY, 14206 39.8 37,946 84.6 22.9 80.0 11.8 7.1
e
Food Lion N/A Charlotte, NC, 28206 33.2 31,617 80.2 29.3 17.5 72.4 11.8
Fred Meyere N/A Portland, OR, 97211 36.0 78,154 94.4 12.1 71.3 16.2 6.8
f
FreshDirect N/A Bronx, NY, 10459 31.3 27,687 65.6 34.3 11.6 34.4 67.9
Giante Peapod Washington, DC, 20017 37.0 77,560 90.1 14.3 29.2 63.3 8.9
e
H-E-B N/A San Antonio, TX, 78227 34.1 40,779 70.2 22.5 91.7 2.4 88.2
JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Hy-Veeef N/A Des Moines, IA, 50316 31.9 44,599 76.4 21.1 63.4 18.6 26.3
e
Kroger N/A Cincinnati, OH, 34205 33.9 63,502 92.6 22.0 64.9 28.6 2.4
Meijere N/A Comstock Park, MI, 49321 36.8 53,686 91.1 10.5 88.5 5.2 8.5
Publixe Instacart Tampa, FL, 33605 35.0 26,954 73.2 35.5 40.2 53.9 27.6
Safewayef N/A Pleasanton, CA, 94566 44.0 153,684 95.9 4.4 64.2 1.9 10.4
ef
ShopRite N/A Hillside, NJ, 07205 37.5 71,102 86.8 10.0 29.2 55.7 19.1
Stop & Shope Peapod Quincy, MA, 02169 38.2 70,248 90.3 11.8 64.8 6.2 3.9
e
Target N/A St. Paul, MN, 55109 37.0 62,858 91.0 10.3 70.4 10.6 6.6
Walmarte,f N/A Bethel Heights, AR, 72764 29.9 47,245 70.6 20.2 61.2 2.4 40.0
Wegmanse Instacart Rochester, NY, 14615 35.8 39,995 84.9 21.7 66.7 23.7 15.0
Wright’s Marketf N/A Auburn, AL, 36832 23.7 27,638 89.3 40.9 63.7 29.6 4.8
a
When different from retailer website.

RESEARCH
b
2014-2018 American Community Survey 5-year estimates.17
c
Among population older than age 25 years.
d
Percent of the population below the federal poverty level.
e
Euromonitor 2018 top-20 revenue-generating retailer (7-11 excluded because classified as convenience store; Trader Joe’s excluded because no online ordering; Walmart Supercenters and Neighborhood Stores included on the Euromonitor list as
two separate retailers but operate through the same e-commerce platform, Walmart.com, which was included in this study).15
f
US Department of Agriculture Supplemental Nutrition Assistance Program Online Purchasing Pilot store (authorized at time of study).16
2297

g
N/A ¼ not applicable (ie, the retailer uses their website as their e-commerce platform).
RESEARCH
2298

Table 2. Description, prevalence, and website location of food and beverage marketing strategies used by 21 US online grocery retailers, 2019-2020
JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Anywhere Department Search Product During


Marketing strategy Description on website Homepage page results pagea pageb checkout

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒn (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Product
Personalized Option to view a customized assortment of 13 (62) N/Ac N/A N/A N/A N/A
storefront products recommended by the retailer
Option to view, shop, Option to allow the retailer to display only 8 (38) N/A N/A N/A N/A N/A
or reorder past previously
purchases purchased products or favorites
Link to loyalty Option to view previous purchases by linking 7 (33) N/A N/A N/A N/A N/A
program online and
in-store accounts
Express shopping list Option to view a list of common products 3 (14) N/A N/A N/A N/A N/A
recommended by
the retailer
Analysis of past Option to allow retailer to analyze past purchases 2 (10) N/A N/A N/A N/A N/A
purchases and make recommendations
Retailer Option to view only products recommended 2 (10) N/A N/A N/A N/A N/A
recommendations specifically for
you by the retailer
Site filters Option to view only products with certain 20 (95) 1 (5) 19 (90) 17 (81)d N/A N/A
characteristics
(eg, brand, nutritional attributes)
Price
Any reduced price Homepage links to a page of products offered at a 21 (100) N/A N/A N/A N/A N/A
December 2022 Volume 122 Number 12

offer reduced price


Coupons Discounts that can be redeemed at checkout 14 (67) N/A N/A N/A N/A N/A
Weekly flyer Link to PDF of flyer showing discounted products 11 (52) N/A N/A N/A N/A N/A
this week
Other time-limited Discounts of a limited duration 11 (52) N/A N/A N/A N/A N/A
savings
Points or vouchers Accrual of rewards that can be redeemed for 6 (29) N/A N/A N/A N/A N/A
savings

(continued on next page)


Table 2. Description, prevalence, and website location of food and beverage marketing strategies used by 21 US online grocery retailers, 2019-2020 (continued)
December 2022 Volume 122 Number 12

Anywhere Department Search Product During


Marketing strategy Description on website Homepage page results pagea pageb checkout

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒn (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Markdowns Instant product markdowns 6 (29) N/A N/A N/A N/A N/A
Loyalty program Savings only for loyalty program members 2 (10) N/A N/A N/A N/A N/A
exclusives
Any personalized Any of the above explicitly described as 8 (38) N/A N/A N/A N/A N/A
reduced price offer personalized or
“made for you”
Placement
Product Products recommended by the retailer that can be 21 (100) 18 (86) 6 (29) 11 (52) 19 (90) 13 (62)
recommendations instantly added to the shopping cart (eg, popular
near you)
Disclosure of Any of the above explicitly described as sponsored, 12 (57) 8 (44) 0 (0) 3 (27) 7 (37) N/A
sponsorship featured,
or paid ads
Any reduced price Any of the above offered at a reduced price 20 (95) 17 (94) 0 (0) 7 (64) 17 (89) N/A
offer
JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Option to sort search Option to sort products based on certain product 19 (90) N/A 18 (86) 16 (76)d N/A N/A
results characteristics
Best match Best match or most relevant items are shown first 17 (89) N/A 4 (22) 13 (81) N/A N/A
by default
Most popular Most popular items are shown first by default 9 (47) N/A 8 (44) 1 (6) N/A N/A
Featured Featured items are shown first by default 5 (26) N/A 3 (14) 2 (13) N/A N/A
Promotion
Title cards Large panel of imagery and text at the top of the 11 (52) 5 (24) 7 (33) 7 (33) 1 (5) N/A
page featuring foods or beverages, typically
below page headers
Any reduced price Title card includes any food or beverage offered at 4 (36) 1 (20) 0 (0) 3 (43) 1 (100) N/A
offer a reduced price
Banner Rectangular Internet display advertising format 12 (57) 4 (19) 4 (19) 8 (38) 4 (19) N/A

RESEARCH
advertisements driving visitors to a post-click landing page,
featuring food or beverages
3 (25) 1 (25) 0 (0) 0 (0) 2 (50) N/A

(continued on next page)


2299
RESEARCH
2300

Table 2. Description, prevalence, and website location of food and beverage marketing strategies used by 21 US online grocery retailers, 2019-2020 (continued)

Anywhere Department Search Product During


JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

Marketing strategy Description on website Homepage page results pagea pageb checkout

Disclosure of Any banner advertisement is explicitly described as


advertising “advertising” or “sponsored”
Any reduced price Any banner advertisement contains a food or 4 (33) 0 (0) 0 (0) 4 (50) 1 (25) N/A
offer beverage offered at a reduced price
Branded product Branded product images are displayed on the site 6 (55) 6 (55) N/A N/A N/A N/A
imagery to illustrate grocery categories or departments
Branded interactive Dynamic site content such as pop-ups, splash 5 (24) 0 (0) 0 (0) 0 (0) 5 (24) N/A
content pages, videos, tutorials, or games contain
branded products
Branded recipes or Instructions to make a meal, recommended foods 11 (52) 1 (5) 1 (5) 0 (0) 4 (19) N/A
meal solutions to eat together as a meal, or meal kits contain
branded products
Product ratings Quantitative customer review (eg, on a 5-star Likert 7 (33) 4 (19) 4 (19) 4 (24) 7 (33) N/A
scale)
Product reviews Qualitative customer review (eg, description of 6 (29) 0 (0) 0 (0) 0 (0) 6 (29) N/A
satisfaction with a product)
Labels Image on or near a product indicating certain 5 (24) 2 (10) 3 (14) 2 (10)d 4 (19) N/A
product characteristics
Calorie labels Calories displayed next to product 2 (40) 0 (0) 1 (33) 0 (0) 2 (50) N/A
Guiding Stars Products are rated 0-3 stars with 0 indicating least 2 (40) 0 (0) 0 (0) 0 (0) 2 (50) N/A
healthy
and 3 indicating healthiest
Warning labels Warnings are displayed next to product to indicate 2 (40) 0 (0) 0 (0) 0 (0) 2 (50) N/A
poor nutritional quality
December 2022 Volume 122 Number 12

Traffic light labels Traffic light labels are displayed next to product to 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) N/A
indicate nutritional quality (red is least healthy,
green is healthiest)
Other nutrition labels Other nutrition labels are displayed next to product 4 (80) 2 (100) 2 (67) 2 (100) 3 (75) N/A
(eg, whole grain, non-GMO,e low-fat)
Link to retailer’s Button to link to retailer’s social media page 14 (67) 14 (67) N/A N/A N/A N/A
social media
Facebook 14 (100) 14 (100) N/A N/A N/A N/A
Twitter 13 (93) 13 (93) N/A N/A N/A N/A

(continued on next page)


Table 2. Description, prevalence, and website location of food and beverage marketing strategies used by 21 US online grocery retailers, 2019-2020 (continued)
December 2022 Volume 122 Number 12

Anywhere Department Search Product During


Marketing strategy Description on website Homepage page results pagea pageb checkout

YouTube 11 (79) 11 (79) N/A N/A N/A N/A


Pinterest 11 (79) 11 (79) N/A N/A N/A N/A
Instagram 8 (57) 8 (57) N/A N/A N/A N/A
Option to interact Button allowing consumer to share opinions about 19 (90) N/A N/A N/A 19 (90) N/A
with the product a product, generate content related to the
product, or move the product to a shopping list
or other saved location
Add to shopping list Button to add the product to a shopping list or 14 (67) N/A N/A N/A 14 (67) N/A
save for later
“Like” product Button to “like” the product 10 (48) N/A N/A N/A 10 (48) N/A
Share product on Button to share product on shopper’s social media 2 (10) N/A N/A N/A 2 (10) N/A
social media page
Upload video Button to upload a video about the product to the 1 (5) N/A N/A N/A 1 (5) N/A
site
a
At each retailer, 15 searches were conducted using the following terms: bread, canned fish, entrée, cheese, chips, canned corn, salad dressing, juice, apples, cookies, pizza, soda, cereal, oil, milk. Values indicate that promotional strategy was present
on at least one search result page unless otherwise noted.
JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS

b
At each retailer, 15 products were added to the cart: Nature’s Own 100% Whole Wheat Bread (20 oz) (Flowers Foods), StarKist Chunk Light Tuna in Water (5 oz) (StarKist Co), Kraft Blue Box Macaroni and Cheese Dinner (7.25 oz) (Kraft Heinz), shredded cheddar
cheese (8 oz, own brand or lowest unit price), Doritos Nacho Cheese Flavored Tortilla Chips (9.75 oz) (PepsiCo), Del Monte Canned Fresh Cut Golden Sweet Kernel Corn (15.25 oz) (Del Monte Pacific Limited), Hidden Valley Ranch Dressing (16 oz) (The Clorox
Company), Tropicana 100% Orange Juice, No Pulp (52 oz) (PepsiCo), Honeycrisp apples, nonorganic (10 count), Double Stuff Oreo (15.35 oz) (Nabisco), Di Giorno Original Rising Crust Pepperoni Pizza (27.5 oz) (Nestlé), Coca-Cola (2 L) (Coca-Cola Company,
Honey Nut Cheerios (15.4 oz) (General Mills), Crisco Pure Canola Oil (48 fl oz) (B&G Foods Inc), 2%/reduced fat cow’s milk (1 gallon, own brand or lowest unit price). Values indicate that promotional strategy was present on at least one product page.
c
N/A ¼ not applicable.
d
Assessed on first search page only (bread).
e
GMO ¼ genetically modified organism.

RESEARCH
2301
RESEARCH

measures for the presence or absence of features on the site. used in the search are listed in in the sequence in which they
Researchers also indicated whether each feature was present were searched. One of each item was added to the cart, with
in specific locations on the site, including the store home additional apples added at the end of the shopping trip to
page, department home page (eg, “breakfast cereal” depart- meet minimum order requirements, when needed. If a spe-
ment), search results page, product page, and at checkout. A cific item was unavailable, researchers selected an item with
detailed description of the measures within each domain is in the same description in a different size or, if an item in a
Table 2. different size was unavailable, researchers selected the most
relevant item in the specific search results (ie, the first item in
the search results list sorted by relevance or best match).
Step 2: Development of Coding Protocol. A protocol for
website coding was developed using an adapted version of Step 3. Expert Review. Before data collection, the Online
McMillan’s framework for conducting web content analyses, Grocery Retail Environment Coding Instrument and website
which includes six steps: drawing a sample for analysis, coding protocol were reviewed and piloted by an expert
identifying a time frame; defining a context unit (ie, section advisory committee, which included members of the Healthy
of the site, such as the home page, that will be coded), Food Retail Working Group—a group of researchers with
defining content categories (ie, features of the content, such expertise in healthy food retail convened by the Nutrition and
as price reductions, that will be coded), training coders and Obesity Policy Research and Evaluation Network.23 The
checking reliability, and analyzing and interpreting the expert advisory committee was asked to complete an
data.19 To capture the complexity and continuous change of abridged version of the shopping activity on a food retail
internet content, websites were coded in real-time rather website of their choice, to code the website using Qualtrics,24
than downloading web content or coding screenshots of and to provide written feedback on their experience. A 60-
static content.20 This approach had limitations, in that two minute teleconference call was held to discuss feedback. On
coders could not independently code each site. However, this call, the two lead researchers asked the expert commit-
coding in real-time accounted for the interactive nature of tee questions about the feasibility of completing the shop-
web content and provided a description of the online food ping task in the retailer they chose, clarity of measures
retail environment that mimics a real-life shopping experi- included in the coding instrument, clarity of the coding
ence. In addition, many features of the online food retail protocol, gaps in the list of measures collected, and general
environment are not revealed unless the consumer interacts impressions of how well the research captured key features
with the site. Coding in real-time was important to capture of the online food retail marketing environment. The re-
the breadth of marketing strategies and features that a con- searchers took notes on this call and used these notes as well
sumer may encounter along the online path to purchase. as feedback provided through Qualtrics24 and via e-mail to
A protocol was developed to prevent researchers’ personal refine the coding instrument and procedures.
information, search histories, and locations from introducing
biased site content. Procedures were developed in consulta-
Step 4. Website Coding. Using the final coding instrument
tion with experts in computer science and involved: con-
and protocol, two graduate-level researchers were randomly
necting to the e-commerce site using a virtual private network
assigned to each code half of the websites in the sample. Data
service, viewing the website in private browsing mode (eg,
were collected at one point in time before the COVID-19
Chrome Incognito [Google LLC], and opening a new browsing
pandemic, from December 12, 2019, to January 2, 2020. Re-
session for each retailer), setting up an account profile at each
searchers were trained by participating in an initial 3-hour
retailer with an anonymous e-mail address and pay-as-you-go
training about the Online Grocery Retail Environment Cod-
mobile telephone number, paying for groceries and fees using
ing Instrument and then independently pilot coding the
anonymous, prepaid debit cards (preferred) or institutional
website of a retailer not included in the study sample. To
credit cards (when prepaid debit not accepted), and delivering
assist with coding, researchers were provided with defini-
groceries to a nearby food pantry.
tions of codebook items, inclusion/exclusion criteria, and
To code each website, researchers completed an activity
examples. The pilot coding was reviewed by a senior member
intended to mimic an authentic online grocery shopping
of the research team and results were discussed to identify
experience. This involved logging into the account profile and
points of disagreement and areas of confusion. This process
completing tasks in the following sequence: navigating to the
was repeated using a different retailer and materials were
grocery home page and select pages linked to the home page
refined until consensus between the two coders was reached.
(eg, weekly flyer); navigating to the breakfast cereal depart-
ment page; searching for a list of 15 generic foods and bev-
erages in random order (eg, “bread”) (Figure 1); searching for Step 5. Assessment of Product Nutritional Quality. While
a specific brand, size, and variety of these 15 foods and coding websites, researchers collected the brand and product
beverages, clicking on each product, and adding each product description of food and beverages marketed using each price,
to the shopping cart; and paying for the order. Throughout placement, and promotion strategy.
the shopping activity, researchers captured images of the site
content. Food categories were selected to reflect a typical and Price. Information for the first five foods and beverages
varied shopping basket, based on the USDA list of most shown on the homepage or linked to the homepage of each
frequently purchased commodity categories.21 Specific food retailer (from Task 1) was recorded: weekly flyer, coupons
items were selected based on Amazon’s list of top-selling (discounts redeemed at checkout), markdowns, loyalty ex-
items by category to ensure availability across stores and clusives (products discounted only for loyalty program
regions.22 The list of food categories, specific items, and terms members), points (products that earn rewards for later

2302 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
RESEARCH

redemption), and other time-limited savings (discounts from the Nutrition Facts label and ingredients list.30 The
offered for a limited time only). PAHO NPM can be used to identify products subject to
regulation (eg, labeling and taxation) and was feasible
to apply based on information from the product label.
Placement. Information for the first five products recom-
Foods and beverages were classified as excessive in so-
mended (ie, products grouped under headings such as
dium (ratio of sodium to energy 1), excessive in free
“popular near you” that were not part of the search results
sugars, which include added sugars and sugars naturally
and could be instantly added to the shopping cart) on the
occurring in honey, syrup, and fruit juices (energy from
homepage, department page, and checkout page, as well as
free sugar 10%), containing other sweeteners (artificial
on each of the 15 generic search results pages and 15 specific
or natural noncaloric sweeteners), and/or excessive in
product pages (the search results themselves were not
saturated fats (energy from saturated fats 10%).
considered recommendations) (from Tasks 1 through 5) was
Because free sugars are not listed on the Nutrition Facts
recorded. On all pages except at the checkout, information for
label, methods for estimating free sugars from total
the first five recommended products that were described as
sugars were used. Like in the PAHO NPM, it was
“sponsored,” and the first five recommended products
assumed that total sugars were equal to free sugars if
offered at a reduced price were also recorded.
the product contained added caloric sweeteners. If the
product was a milk, yogurt, or processed fruit with
Promotion. Information for the first five products shown in added caloric sweeteners, free sugars were estimated as
title cards (large panels of imagery and text at the top of the half the total sugar content of the product.
page), banner advertisements (rectangular advertisements - Products were also classified into 13 food groups
driving visitors to a postclick landing page), interactive site based on a coding scheme developed by Franckle and
content (eg, videos and games), and recipes on the home colleagues.31 A detailed description of items included
page, department page, search results pages, and product in each group is available in Figure 2 (available at
pages (from Tasks 1 through 4) was recorded. Researchers www.jandonline.org).
also recorded information for the first five products shown in
title cards and banner advertisements at a reduced price, and Data Analysis
recorded information for branded product imagery used to First, the proportion of retailers using each marketing strat-
illustrate grocery categories on the homepage. On each egy anywhere on the website and at different points along
generic search results page, the brand and product descrip- the path to purchase (ie, home page, department page,
tion for the first five products in the list of search results was search results page, product page, and checkout) was
recorded to assess the degree to which results matched the tabulated. Second, the proportion of marketed products
intent of the search (eg, does a search for chicken return that fell into each of 13 food groups was tabulated. Third,
chicken nuggets or fresh poultry). the proportion of marketed products that were ultra-
For each product recorded, researchers entered informa- processed and of poor nutritional quality overall and by
tion from the Nutrition Facts label and ingredients list into a marketing strategy was calculated. Data were analyzed
database using information from the store website. Products using Stata SE version 15.1.32
were defined as “poor nutritional quality” in the case that
they were ultraprocessed and were excessive in key nutri- Human Subjects Protections
ents. Procedures for determining nutritional quality are
This study was deemed nonhuman subjects research by the
described in detail below.
Johns Hopkins Bloomberg School of Public Health Institu-
- Two researchers categorized food and beverages as tional Review Board.
ultra-processed or not using the NOVA classification
system.25 A third researcher independently coded all RESULTS
products and discrepancies were resolved by the
principal investigator. Products were assigned to NOVA Retailer Characteristics
category 4 (ultraprocessed) or NOVA categories 1-3 Retailers (N ¼ 21) were located in all nine US census regions
(not ultraprocessed) by reviewing the product’s in- in areas serving socioeconomically, racially, and ethnically
gredients list. More information about this methodol- diverse populations (Table 1). Median household income
ogy has been published previously.25 The decision to within the Zip code of each store’s location ranged from
classify products as ultra-processed or not was based $26,954 (Publix) to $153,684 (Safeway) and percent poverty
on global evidence that high consumption of ultra- ranged from 4.4% (Safeway) to 40.9% (Wright’s Market). The
processed foods (ie, NOVA 4) is associated with proportion of the population identifying as Black ranged from
increased risk of poor health outcomes, including 1.6% (Albertson’s) to 72.4% (Food Lion) and the Hispanic
weight gain and related comorbidities, in children and population ranged from 2.4% (Kroger) to 88.2% (H-E-B).
adults.26 There have been several global calls to action
to reduce consumption of these products.27,28 Marketing Strategies
- Because there is variation in nutritional quality among Product. Two marketing strategies intended to alter the
foods even within the ultraprocessed foods category,29 product mix available to consumers were identified:
the Pan-American Health Organization (PAHO) Nutrient personalized storefronts and website filters (see Table 2 for a
Profile Model (NPM) was used to further classify foods glossary of terms). Personalized storefronts allowed con-
according to nutrient density based on information sumers to opt-in to one of several options to view a

December 2022 Volume 122 Number 12 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2303
RESEARCH

customized assortment of products recommended by the based rating systems such as Guiding Stars (40%), calorie la-
retailer. Thirteen (62%) retailers offered this option, either by bels (40%), or traffic light labels (0%).
displaying products previously ordered online or “favorites,”
linking online and in-store accounts to display past pur- Nutritional Quality of Marketed Products
chases, displaying a list of “commonly purchased” or “pop-
When the nutritional quality of the first five foods or bever-
ular” products, analyzing past purchases and making
ages marketed across all retailers was examined, 64% were
personalized recommendations, or displaying only products
found to be ultraprocessed (NOVA category 4); 36% were not
recommended “for you” by the retailer. Nearly all retailers
ultraprocessed (NOVA categories 1 through 3); and 62% met
(95%) offered website filters, or the option to view only
the criteria for “poor nutritional quality” (ultraprocessed and
products with certain characteristics (eg, brand). Filters were
exceeding at least one PAHO criterion) (Table 3). Of the
infrequently offered on the home page (5%) but were often
ultraprocessed foods, 69% were excessive in sodium, 49%
offered on the department page (90%) or search results page
were excessive in free sugars, 40% were excessive in saturated
(81%).
fats, and 4% contained low- or no-calorie sweeteners. Prod-
ucts marketed through banner advertisements (87%) and
Price. All retailers offered products at a reduced price, either
sponsored product recommendations (76% on the home
through digital coupons (67%), a weekly flyer (52%), time-
page, 71% on the search results page, and 66% on the product
limited savings (52%), rewards (29%), instant product mark-
page) were more likely to be of poor nutritional quality
downs (29%), and/or loyalty program exclusives (10%). Eight
compared with products marketed through other means.
retailers (38%) offered price reductions that were explicitly
When marketed products were examined by food group, the
personalized (ie, described as made “for you”). Price re-
top-five marketed categories were: candy, sweets, and snacks
ductions were often paired with other strategies, including
(17.3%); fruit, vegetables, and legumes (16.7%); bread, cereal,
product recommendations (95%), title cards (36%), and ban-
and grains (10.8%); milk, yogurt, and cheese (9.7%); and
ner advertisements (33%).
sugary drinks (6.9%) (Table 4). There was some variation by
marketing type, with bread, cereal, and grains accounting for
Placement. All retailers offered recommendations for the largest proportion of price reductions (18.4%); fruit,
products that could be instantly added to the cart. This
vegetables, and legumes accounting for the largest propor-
strategy was most common on the home page (86%) and
tion of product recommendations (17.1%); and candy, sweets,
product page (90%) but was also prevalent on the department
and snacks accounting for the largest proportion of other
page (29%), search results page (52%), and at checkout (62%).
promotions (36.4%). When searching for foods, the nutri-
Recommended products were grouped under themes, which,
tional quality of the first five search results varied by cate-
in some cases, were explicitly personalized (eg, “popular near
gory. Although few results for oil (1%) (eg, oil-based salad
you,” “you might also like,” or “recommended for you”). Just
dressing), apples (2%) (eg, sliced apples with caramel dip),
more than half of retailers in the sample (57%) disclosed
and milk (11%) (eg, chocolate milk) were of poor nutritional
when product recommendations were sponsored; for others,
quality, most results for cereal (99%), salad dressing (97%),
sponsorship was unknown. Nearly all retailers (90%) allowed
and bread (95%) were unhealthy (ie, ultraprocessed and
customers to sort products by certain characteristics, such as
excessive in key nutrients to limit) (Table 5, available at www.
price or nutritional quality. By default, products were most
jandonline.org).
commonly sorted by relevance (89% of retailers offering a sort
option), popularity (47%), or featured items (26%).
DISCUSSION
Promotion. A wide variety of promotional strategies were This research examined the volume and content of food and
used by retailers to increase visibility or appeal of certain beverage marketing, as well as the nutritional quality of
products. Most retailers (52%) used title cards to feature food marketed products, in 21 US online grocery stores, including
or beverages on the home page (24%), department page major chain retailers and those participating in the SNAP
(33%), search results page (33%), and/or product page (5%). Online Purchasing Pilot.
Banner advertisements were also popular (57%); only three A wide variety of marketing strategies were employed,
retailers explicitly described these as sponsored advertise- individually and in combination, across the online grocery
ments. Some retailers used images of specific, branded stores surveyed. These included the strategic placement of
products to designate grocery categories or departments certain products, and strategies to alter the assortment of
(55%), in dynamic site content such as pop-ups, splash pages, products, reduce their price, and increase their visibility and
videos, and/or tutorials (24%), or in recommended recipes appeal. Most retailers offered personalized storefronts and
(52%). Some retailers enabled peer-to-peer communication, nearly all enabled filtering of web pages by product charac-
including product ratings (33%) and reviews (29%). Most re- teristics, such as brand, price, or nutrition-related attributes.
tailers enabled customer interaction with products (90%) by All 21 retailers offered products online at a reduced price—a
adding products to a shopping list, “liking” a product, sharing pricing strategy that was frequently combined with other
a product on social media, or uploading a product video re- marketing strategies, such as product recommendations and
view to the site. Most retailers provided links to their own banner advertisements. Retailers used product recommen-
social media pages (67%) such as Facebook or Twitter. Food dations and search result ordering to increase consumers’
labels were relatively unpopular (24%). Of retailers displaying exposure to certain products and brands. Some of these
labels, most (80%) used their own labeling system to high- recommendations were explicitly personalized (eg, “made for
light nutrition-related attributes of products (eg, whole grain, you”), or paid product placements (eg, “sponsored” prod-
nongenetically modified, or low-fat); fewer used evidence- ucts), and others were likely, but not explicitly, paid (eg,

2304 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
RESEARCH

Table 3. Nutritional quality of food and beverages marketed by 21 US online grocery retailers, 2019-2020, by marketing strategy
(n ¼ 3,473)a

Total Poor Contains Excessive


products nutritional Ultra-processed Excessive Excessive other saturated
Marketing strategy assessed qualityb product sodium free sugars sweeteners fats

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒn (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Overall 3,473 2,155 (62) 2,214 (64) 1,519 (69) 1094 (49) 95 (4) 877 (40)
Price
All reduced price offers 212 123 (58) 127 (60) 86 (68) 85 (67) 5 (4) 51 (40)
Weekly flyer 47 19 (40) 20 (43) 12 (60) 12 (60) 1 (5) 10 (50)
Other time-limited 52 30 (58) 30 (58) 22 (73) 20 (67) 0 (0) 13 (43)
savings
Coupons 65 45 (69) 47 (72) 29 (62) 33 (70) 1 (2) 21 (45)
Markdowns 16 12 (75) 13 (81) 9 (69) 8 (62) 3 (23) 2 (15)
Loyalty program 23 15 (65) 15 (65) 12 (80) 12 (80) 0 (0) 3 (20)
exclusives
Points or vouchers 9 2 (22) 2 (22) 0 (0) 0 (0) 0 (0) 2 (100)
Placement
All product 2,744 1,659 (60) 1,704 (62) 1,252 (73) 767 (45) 77 (5) 619 (36)
recommendationsc
Homepage product 100 61 (61) 62 (62) 39 (63) 36 (58) 8 (13) 26 (42)
recommendations
Sponsored 17 13 (76) 13 (76) 10 (77) 4 (31) 2 (15) 5 (38)
Reduced price 74 45 (61) 46 (62) 24 (52) 29 (63) 6 (13) 19 (41)
Department page 25 12 (48) 14 (56) 7 (50) 8 (57) 0 (0) 1 (7)
product
recommendations
Sponsored 0 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0)
Reduced price 7 3 (43) 5 (71) 3 (60) 2 (40) 0 (0) 0 (0)
Search results page 711 377 (53) 388 (55) 286 (74) 195 (50) 16 (4) 119 (31)
product
recommendations
Sponsored 139 99 (71) 106 (76) 59 (56) 56 (53) 8 (8) 35 (33)
Reduced price 172 98 (57) 99 (58) 86 (87) 53 (54) 5 (5) 32 (32)
Product page 1,202 744 (62) 760 (63) 568 (75) 306 (40) 28 (4) 294 (37)
recommendations
Sponsored 109 72 (66) 77 (71) 57 (74) 26 (34) 12 (16) 22 (29)
Reduced price 404 295 (73) 306 (76) 221 (72) 126 (41) 14 (5) 118 (39)
Checkout page product 49 24 (49) 24 (49) 18 (75) 12 (50) 0 (0) 10 (42)
recommendations
Promotion
All other promotions 517 373 (72) 383 (74) 181 (47) 242 (63) 13 (3) 207 (54)
Title cards 104 54 (52) 57 (55) 38 (67) 31 (54) 3 (5) 15 (26)
Reduced price 42 15 (36) 16 (38) 11 (69) 11 (69) 0 (0) 6 (38)
Banner advertisements 299 259 (87) 261 (87) 99 (38) 186 (71) 6 (2) 174 (67)
Reduced price 16 16 (100) 16 (100) 16 (100) 3 (19) 2 (13) 2 (13)
(continued on next page)

December 2022 Volume 122 Number 12 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2305
RESEARCH

Table 3. Nutritional quality of food and beverages marketed by 21 US online grocery retailers, 2019-2020, by marketing strategy
(n ¼ 3,473)a (continued)

Total Poor Contains Excessive


products nutritional Ultra-processed Excessive Excessive other saturated
Marketing strategy assessed qualityb product sodium free sugars sweeteners fats

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒn (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Branded product imagery 23 8 (35) 10 (43) 5 (50) 4 (40) 0 (0) 3 (30)
Branded interactive 21 14 (67) 16 (76) 7 (44) 5 (31) 2 (13) 5 (31)
content
Branded recipes 12 7 (58) 7 (58) 5 (71) 2 (29) 0 (0) 2 (29)
a
Dataset of products compiled by recording nutrition information for the first five foods and beverages marketed using each marketing strategy on each webpage of 21 grocery retailers.
b
Products are defined as of poor nutritional quality in the case that they are ultra-processed and excessive in sodium, excessive in free sugars, ontaining other sweeteners, or excessive in
saturated fats. Values are presented as % of ultraprocessed product.
c
The search results themselves were not included.

“featured” products). Other promotional strategies varied pages did not meet the National Alliance on Nutrition and
widely by retailer and included banner advertisements (some Activity (a coalition that advocates for healthy eating policies)
of which were labeled “sponsored”), peer-to-peer commu- nutrition guidelines.9 One study of cereal marketing on
nication, branded content, interactive product features, user- manufacturer websites targeting children found that 64% of
generated content, and social media engagement. advertised products fell short of the National Academy of
This is one of the first studies to examine the types of Medicine guidelines for competitive school foods.36 One
marketing strategies or the prevalence of their use within reason for these differences in marketing by food category
online food retail platforms in the United States. Previous may be that consumers are concerned about the quality of
work has suggested that product price promotions, front-of- fresh and refrigerated foods and prefer to purchase these
package labeling, and availability of nutrition information on items in-person—a trend that may change as customers
packaged foods differ between online and physical super- become more comfortable shopping for groceries online.37 In
markets.14,33e35 For example, a study of six UK supermarkets addition, compared with producers of fresh foods, manufac-
found that front-of-package labels were present on 74% of turers of packaged foods and beverages have larger market-
products from physical stores, compared with 42% of iden- ing budgets to pay high trade fees for prime placement in the
tical products in online stores.14 By contrast, more price store.5,6,38 Retailers have also cited difficulty cross-promoting
promotions were available in physical supermarkets (32% of fresh foods due to structural limitations (eg, lack of refriger-
products) compared with online stores (24%). Three studies ation), perishability, and coordination across departments (ie,
surveying online grocery retailers in the United States found produce and grocery are often run by different category
that many packaged products lacked conspicuous, legible managers).39 Although information about the fees paid for
nutrition information required on all packaged foods sold in product placement in online stores is proprietary, the online
the United States, including Nutrition Facts labels, allergen setting can overcome many of the structural constraints of
information, and ingredients lists.33-35 Researchers have the brick-and-mortar store. Indeed, in this study, the fruit,
found that federal agencies overseeing food labeling and vegetable, and legume category accounted for 16.7% of mar-
online advertising, such as the Food and Drug Administration keted products overall and was the category most likely to be
and Federal Trade Commission, have authority to require that recommended to consumers along the path to purchase, ac-
online food retailers alter their practices in support of counting for 17.1% of product recommendations.
nutrition equity and informed consumer decision making This research is subject to several limitations and strengths.
online.35 Congress and USDA could also leverage SNAP- First, despite using multiple methods to ensure researchers’
authorized retailers to set rules about the display of nutri- personal information and browsing histories did not influ-
tional information, product recommendations, or other ence product display on retailer websites, we cannot know if
product marketing techniques. this was entirely prevented. Second, websites were captured
Of nearly 3,500 products examined in this study, two- at one point in time and observed product marketing may
thirds were ultraprocessed (64%) and of poor nutritional have been influenced by seasonal marketing trends. Third,
quality (62%). Candy, sweets, and snacks made up the largest the types of products purchased and the order in which
percentage of food and beverage promotions (17.3%), fol- products were searched may have influenced marketing on
lowed by fruit, vegetables, and legumes (16.7%). That most the site; however, the goal of the study was to understand
food and beverages marketed in online grocery stores are of what a typical consumer might see; efforts were made to
poor nutritional quality is not surprising. A report from select a representative basket of food and beverages, and
Center for Science in the Public Interest, a national nutrition items were selected in random order. Fourth, researchers
advocacy group, examined product marketing in six online made a one-time purchase of 15 items at each store; we
grocery retailers and found that, on average, half (51%) of cannot comment on marketing strategies that may be
products marketed on retailer home pages and search result revealed or become more sophisticated with repeat

2306 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
RESEARCH

Table 4. Food groups marketed by 21 US online grocery retailers, 2019-2020a

All marketing Price Placement Promotion


All price, placement, and All reduced-price All product All other
Food group promotional strategies offers recommendations promotions

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒn (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ!
Total 3,473 (100) 212 (100) 2,744 (100) 517 (100)
Candy, sweets, and snacks 602 (17.3) 37 (17.5) 377 (13.7) 188 (36.4)
Fruit, vegetables, and 581 (16.7) 38 (17.9) 469 (17.1) 74 (14.3)
legumes
Bread, cereal, and other 472 (13.6) 39 (18.4) 397 (14.5) 36 (7.0)
grains
Mixed dishes and prepared 375 (10.8) 17 (8.0) 318 (11.6) 40 (7.7)
foods
Milk, yogurt, and cheese 338 (9.7) 10 (4.7) 276 (10.1) 52 (10.1)
Sugary drinks 241 (6.9) 14 (6.6) 189 (6.9) 38 (7.4)
Seafood, meat, poultry, and 195 (5.6) 30 (14.2) 142 (5.2) 23 (4.5)
eggs
Condiments and sauce 167 (4.8) 5 (2.4) 154 (5.6) 8 (1.6)
Unsweetened drinks 163 (4.7) 7 (3.3) 137 (5.0) 19 (3.7)
100% juice 106 (3.1) 0 (0) 100 (3.6) 6 (1.2)
Low-calorie drinksb 79 (2.3) 4 (1.9) 63 (2.3) 12 (2.3)
Fats and oils 77 (2.2) 1 (0.5) 68 (2.5) 8 (1.6)
Otherc 77 (2.2) 10 (4.7) 54 (2.0) 13 (2.5)
a
Dataset of products compiled by recording nutrition information for the first five foods and beverages marketed using each marketing strategy on each webpage of 21 grocery retailers.
b
Includes drinks sweetened with low- or no-calorie sweeteners.
c
Includes alcohol, processed meat and soy, and other cooking ingredients (eg, flour and sugar).

purchases. Fifth, we did not capture personalized marketing than 1,000 unique pages across a diverse group of online
unless personalization was explicit. Although personalization grocery retail websites, use of a validated conceptual frame-
is an important characteristic of online food retail marketing, work specific to online food retail in codebook development,
this construct was not captured in this study because there assessment of the nutritional quality of marketed products
are many factors that may contribute to variation between using several methods, and the development of a robust
sites other than dimensions that researchers intentionally methodology for assessing marketing in online retail.
vary, such as geolocation or prior purchases. For example,
two search queries from the same IP address may be directed
to different data centers, resulting in variation in site content PUBLIC HEALTH IMPLICATIONS
that is not caused by intentional personalization.40 Other Online grocery has the potential to address healthy food access
algorithms, A/B testing on the site, or unknown site archi- and support healthy purchases. One randomized trial in a food
tecture may also cause variation in site content. Thus, pantry offering online grocery shopping found that a prefilled,
personalization was deemed outside the scope of this study. nutritionally balanced online shopping cart increased pur-
Sixth, this study did not assess how marketing affects con- chases of fruit, vegetables, and whole grains compared with
sumer behavior, or if the decision-making process differed nutrition education alone.41 A randomized crossover trial of-
online vs in-store. Lastly, these data were collected before the fering personalized price reductions for targeted healthful
COVID-19 pandemic, during which time the online grocery foods saw a small increase in purchases of these foods over 6
sector grew substantially. New marketing practices may be months compared with a control condition in which price re-
introduced as grocery e-commerce sites become more so- ductions were not personalized or targeted.42 Another ran-
phisticated and as new, third-party vendors enter the market. domized trial in an experimental online supermarket in the
This research is the first to capture the variety and relative United Kingdom found that providing a shopping list in which
volume of marketing strategies used at different points along items were ranked in ascending order of saturated fat content
the path to purchase in 21 US grocery chains. The stores reduced energy from saturated fat in food purchases by 5%
included in the research were located in areas serving soci- compared with no intervention.43 Although these results are
odemographically diverse populations, increasing generaliz- promising, they may be threatened by the marketing of un-
ability of results. Other strengths include analysis of more healthful products. No study, to our knowledge, has assessed

December 2022 Volume 122 Number 12 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2307
RESEARCH

Product Generic
category Product description, size search terms Specific search terms
a
Bread Nature’s Own 100% Whole Wheat, 20 oz (Flowers Foods; Bread Nature’s Own Whole
Thomasville, GA) Wheat Bread
Canned fish Starkist Chunk Light Tuna in Water,b 5 oz (Starkist Co; Reston, VA) Canned fish StarKist Chunk Light
Tuna
Mixed dish, Kraft Blue Box Macaroni and Cheese Dinner,c 7.25 oz (Kraft Heinz Entrée Kraft Macaroni and
entrée Co; East Hanover, NJ) Cheese
Cheese Generic brand shredded cheddar, 8 oz Cheese Shredded cheddar
d
Salty snacks Doritos Nacho Cheese Flavored Tortilla Chips, 9.75 oz (PepsiCo; Chips Doritos
New Brunswick, NJ)
Canned Del Monte Canned Fresh Cut Golden Sweet Kernel Corn,e 15.25 oz Canned corn Del Monte Corn
vegetables (Del Monte Foods; Bloomfield, NJ)
Salad Hidden Valley Ranch Dressing,f 16 fl oz (Hidden Valley; Fullerton, Salad Hidden Valley Ranch
dressing CA) dressing
Juice Tropicana Orange Juice 100% Juice, No Pulp,d 52 fl oz (PepsiCo) Juice Tropicana Orange
Juice
Fresh fruit Honeycrisp apples, conventionally grown, 10 count Apples Honeycrisp apples
Cookies Nabisco Double Stuf Oreo,g 15.35 oz (Nabisco; East Hanover, NJ) Cookies Double Stuf Oreo
h
Frozen pizza DiGiorno Original Rising Crust Pepperoni Pizza, Pizza DiGiorno Pizza
27.5 oz (Nestle; Vevey, Switzerland)
Soda Coca-Cola,i 2 L (Coca-Cola Co; Monmouth Junction, NJ) Soda Coke
j
Cereal General Mills Honey Nut Cheerios, 15.4 oz (General Mills Inc; Cereal Honey Nut Cheerios
Dresher, PA)
Fats and oils Crisco Pure Canola Oil,k 48 fl oz (B&G Foods; Parsippany, NJ) Oil Crisco Canola Oil
Cow’s milk Generic brand 2% reduced fat milk, unflavored, nonorganic, 1 gal Milk 2% milk
a
Flowers Foods.
b
StarKist Co.
c
Kraft Heinz.
d
PepsiCo.
e
Del Monte Pacific Limited.
f
The Clorox Company.
g
Nabisco.
h
Nestlé.
i
The Coca-Cola Company.
j
General Mills.
k
B&G Foods Inc.
Figure 1. Description of products included in the grocery basket and search terms used in the online grocery shopping task

how retailer marketing practices influence consumer food If online grocery marketing affects dietary behaviors,
choices in real-world e-commerce settings. Research of this personalized marketing may exacerbate longstanding health
nature could inform policy changes that shift marketing to- inequities. A recent report highlights how Black, Latinx, and
ward more healthful products, such as marketing re- low-income individuals may become more reliant on online
quirements for SNAP-authorized retailers. retailers to overcome poor access to healthful foods in their

2308 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
RESEARCH

communities.11 This reliance exposes individuals to greater 12. Mathur A, Acar G, Friedman MJ, et al. Dark patterns at scale: findings
collection of personal data and predatory marketing of foods from a crawl of 11K shopping websites. Proc ACM Hum Comput
Interact. 2019;3(81):1-32. https://doi.org/10.1145/3359183
and beverages based on presumed preferences and past be-
13. Chester J, Montgomery KC, Kopp K. Big food, big tech, and the global
haviors. The types of products and services advertised are childhood obesity pandemic. May 12, 2021. Accessed May 26, 2021.
informed by algorithms, which may be rooted in systemic bia- https://www.democraticmedia.org/article/big-food-big-tech-and-
ses that disproportionately influence already disadvantaged global-childhood-obesity-pandemic
groups and can entrench existing inequities.44 More research is 14. Bhatnagar P, Scarborough P, Kaur A, Dikmen D, Adhikari V,
Harrington R. Are food and drink available in online and physical
needed to understand the degree to which the marketing supermarkets the same? A comparison of product availability,
practices identified in this study are personalized, and the in- price, price promotions and nutritional information. Public Health
fluence of such personalization on the dietary behaviors of Nutr. 2021;24(5):819-825. https://doi.org/10.1017/S136898002000
4346
populations at highest risk of diet-related chronic diseases.
15. Euromonitor International. Top U.S. supermarkets by revenue.
February 2021. Accessed May 26, 2021. https://www.euromonitor.
com/supermarkets-in-the-us/report#
CONCLUSIONS 16. US Dept of Agriculture. USDA launches SNAP Online Purchasing Pi-
Online grocery retailers commonly use a variety of strategies— lot: Participants may buy groceries online in New York. April 18,
2019. Accessed May 26, 2021. https://www.fns.usda.gov/pressrelea
personalized storefronts, price reductions, product recom- se/2019/fns-000319
mendations, search result ordering, banner advertisements, 17. US Census Bureau. American Community Survey 2014-2018 5-year
branded site content, and interactive media—to market food estimates now available. December 19, 2019. Updated October 8,
and beverages on their websites. This marketing is prominent 2021. Accessed May 26, 2021. https://www.census.gov/newsroom/
press-releases/2019/acs-5-year.html
on the home page, search results page, department page,
18. Moran A, Perez C, Headrick G Taillie LS, Khandpur N. Capturing
product page, and at checkout and predominantly features
marketing and policy practices of online food retailers: a codebook.
food and beverages of poor nutritional quality. Future research Accessed April 21, 2022. https://doi.org/10.17605.OSF.IO/VY9T
should examine the extent to which this marketing is 19. McMillan SJ. The microscope and the moving target: the challenge of
personalized on the basis of consumer characteristics, and its applying content analysis to the world wide web. Journal Mass
influence on food choices and dietary behaviors. Commun Q. 2001;77(1). https://doi.org/10.1177/1077699000077
00107
20. Kim I, Kuljis J. Applying content analysis to web-based content.
References J Comput Inf Technol. 2010;18(4):369-375.
1. Redman R. Online grocery to more than double market share by
2025. Supermarket News. September 18, 2020. Accessed May 26, 21. Garasky S, Mbwana K, Romualdo A, Tenaglio A, Roy M. Foods typi-
2021. https://www.supermarketnews.com/online-retail/online- cally purchased by SNAP households: Appendices. November 2016.
grocery-more-double-market-share-2025 Accessed May 26, 2021. https://fns-prod.azureedge.net/sites/default/
files/ops/SNAPFoodsTypicallyPurchased.pdf
2. 2020 online grocery shopping statistics: pre and post covid-19.
Superfood.com. April 16, 2020. Accessed May 26, 2021. https:// 22. Best sellers grocery gourmet food. Updated November 19; 2021.
superfood.digital/online-grocery-store-ecommerce-statistics/ Accessed September 1, 2019. https://www.amazon.com/Best-
Sellers-Grocery-Gourmet-Food/zgbs/grocery. 2021
3. Prime is just $5.99/month* for qualifying government assistance
recipients. Accessed May 26, 2021. https://www.amazon.com/58f8 23. Nutrition & Obesity Policy Research and Evaluation Network (NOP-
026f-0658-47d0-9752-f6fa2c69b2e2/qualify REN). Healthy Retail Work Group. Accessed May 26, 2021. https://
nopren.org/working_groups/healthy-food-retail/
4. Moran A, Headrick G, Khandpur N. Promoting equitable expansion of
the SNAP Online Purchasing Pilot. July 2021. Accessed November 30, 24. Qualtrics. Qualtrics; 2020.
2021. https://healthyeatingresearch.org/wp-content/uploads/2021/ 25. Monteiro CA, Cannon G, Levy RB, et al. Ultra-processed foods: what
07/HER-Online-Purchasing-Brief-FINAL.pdf they are and how to identify them. Public Health Nutr. 2019;22(5):
5. Rivlin G. Rigged: Supermarket shelves for sale. September 2016. 936-941. https://doi.org/10.1017/S1368980018003762
Accessed May 26, 2021. https://cspinet.org/sites/default/files/ 26. Crimarco A, Landry MJ, Gardner CD. Ultra-processed foods, weight
attachment/Rigged%20report_0.pdf gain, and co-morbidity risk. Curr Obes Rep. 2021:1-13. https://doi.
6. Hecht AA, Perez CL, Polacsek M, Thorndike AN, Franckle RL, org/10.1007/s13679-021-00460-y
Moran AJ. Influence of food and beverage companies on retailer 27. Popkin BM, Barquera S, Corvalan C, et al. Towards unified and im-
marketing strategies and consumer behavior. Int J Environ Res Public pactful policies to reduce ultra-processed food consumption and
Health. 2020;17(20):7381. promote healthier eating. Lancet Diabetes Endocrinol. 2021;9(7):462-
7. Glanz K, Bader MD, Iyer S. Retail grocery store marketing strategies 470. https://doi.org/10.1016/S2213-8587(21)00078-4
and obesity: an integrative review. Am J Prev Med. 2012;42(5):503- 28. Khandpur N, Neri DA, Monteiro C, et al. Ultra-processed food
512. https://doi.org/10.1016/j.amepre.2012.01.013 consumption among the paediatric population: an overview and
8. Khandpur N, Zatz L, Bleich S, et al. Supermarkets in cyberspace: a call to action from the European Childhood Obesity Group. Ann
conceptual framework to capture the influence of online food retail Nutr Metab. 2020;76(2):109-113. https://doi.org/10.1159/00050
environments on consumer behavior. Int J Environ Res Public Health. 7840
2020;17(22):8639. https://doi.org/10.3390/ijerph17228639 29. Mozaffarian D, El-Abbadi NH, O’Hearn M, et al. Food Compass is a
9. McCarthy J, Minovi D, Wootan M. Scroll and shop: Food marketing nutrient profiling system using expanded characteristics for assess-
migrates online. January 2020. Accessed June 19, 2020. https:// ing healthfulness of foods. Nat Food. 2021;2:809-818. https://doi.org/
cspinet.org/sites/default/files/attachment/Scroll_and_Shop_report. 10.1038/s43016-021-00381-y
pdf 30. Pan American Health Organization. Nutrient Profile Model; 2016.
10. Jilcott Pitts SB, Ng SW, Blitstein JL, Gustafson A, Niculescu M. Online Accessed October 26, 2021. https://www.paho.org/en/nutrient-
grocery shopping: promise and pitfalls for healthier food and profile-model.
beverage purchases. Public Health Nutr. 2018;21(18):3360-3376. 31. Franckle RL, Moran A, Hou T, et al. Transactions at a northeastern
https://doi.org/10.1017/S1368980018002409 supermarket chain: differences by Supplemental Nutrition Assis-
11. Chester J, Kopp K, Montgomery KC. Does buying groceries online put tance Program use. Am J Prev Med. 2020;59(2):305. https://doi.org/
SNAP participants at risk? How to protect health, privacy, and equity. 10.1016/j.amepre.2017.06.019
July 2020. Accessed May 26, 2021. https://www.democraticmedia.org/ 32. Stata Statistical Software: Release 15. StataCorp LP; 2017. www.stata.
sites/default/files/field/public-files/2020/cdd_snap_report_ff_0.pdf com

December 2022 Volume 122 Number 12 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2309
RESEARCH

33. Headrick G, Khandpur N, Perez C, Taillie LS, Bleich SN, Rimm EB, 39. Greene J. Understanding the value of academic research partner-
Moran A. Content analysis of online grocery retail policies and ships with food retailers. December 2020. Accessed May 26, 2021.
practices affecting healthy food access. J Nutr Educ Behav. https://healthyeatingresearch.org/wp-content/uploads/2020/11/HER-
2022;54(3):219-229. Academic-Research-Partnerships-3.pdf
34. Olzenak K, Fresnch S, Sherwood N, Redden JP, Harnack L. How online 40. Hannak A, Soeller G, Lazer D, Mislove A, Wilson C. Measuring price
stores support consumer nutrition information needs. J Nutr Educ Behav. discrimination and steering on e-commerce web sites. Paper pre-
2020;53(10):952-957. https://doi.org/10.1016/j.jneb.2020.07.009 sented at: IMC ’14: Proceedings of the 2014 Conference on Internet
Measurement. Vancouver, BC, Canada, November 5-7, 2014. https://
35. Pomeranz JL, Cash SB, Springer M, Del Guidice IM, Mozaffarian D. doi.org/10.1145/2663716.2663744
Opportunities to address the failure of online food retailers to
ensure access to required food labelling information in the USA. 41. Coffino JA, Udo T, Hormes JM. Nudging while online grocery shop-
Public Health Nutr. 2022:1-9. https://doi.org/10.1017/S136898002 ping: a randomized feasibility trial to enhance nutrition in in-
1004638 dividuals with food insecurity. Appetite. 2020;152(1):104714.
https://doi.org/10.1016/j.appet.2020.104714
36. Cheyne AD, Dorfman L, Bukofzer E, Harris JL. Marketing sugary ce-
reals to children in the digital age: a content analysis of 17 child- 42. Vadiveloo M, Guan X, Parker HW, et al. Effect of personalized in-
targeted websites. J Health Commun. 2013;18(5):563-582. https:// centives on dietary quality of groceries purchased: a randomized
doi.org/10.1080/10810730.2012.743622 crossover trial. JAMA Network Open. 2021;4(2):e2030921. https://doi.
org/10.1001/jamanetworkopen.2020.30921
37. NielsenIQ. Which health and food segments are missing out on
the e-commerce boom? March 9, 2021. Accessed on May 26, 43. Koutoukidis DA, Jebb SA, Ordonez-Mena JM, et al. Prominent
2021, https://nielseniq.com/global/en/insights/analysis/2021/key- positioning and food swaps are effective interventions to reduce
consumer-needs-fuel-in-store-sales-during-covid-19-e-commerce- the saturated fat content of the shopping basket in an experimental
boom/ online supermarket: a randomized controlled trial. Int J Behav Nutr
Phys Act. 2019;16(1):50. https://doi.org/10.1186/s12966-019-
38. Karpyn A, McCallops K, Wolgast H, Glanz K. Improving consumption
0810-9
and purchases of healthier foods in retail environments: a systematic
review. Int J Environ Res Public Health. 2020;17(20):7524. https://doi. 44. Eubanks V. Automating Inequality: How High-Tech Tools Profile, Police,
org/10.3390/ijerph17207524 and Punish the Poor. St Martin’s Press; 2018.

AUTHOR INFORMATION
A. J. Moran is an assistant professor and C. Perez is a doctoral degree candidate, Department of Health Policy and Management, and G. Headrick
is a doctoral degree candidate, Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD. A.
Greatsinger is a senior health care analyst, Analysis Group, Boston, MA. L. S. Taillie is an assistant professor, Department of Nutrition, Gillings
School of Global Public Health, Carolina Population Center, University of North Carolina, Chapel Hill. L. Zatz is a senior advisor, The Behavioural
Insights Team North America, Washington, DC; at the time of the study, she was a doctoral degree candidate, Department of Nutrition, Harvard
TH Chan School of Public Health, Boston, MA. S. N. Bleich is a senior advisor for COVID, Office of the Secretary, US Department of Agriculture,
Washington, DC; at the time of the study, she was a professor, Department of Health Policy and Management, Harvard TH Chan School of Public
Health, Boston, MA. E. B. Rimm is a professor, Departments of Nutrition and Epidemiology, Harvard TH Chan School of Public Health, Boston, MA.
N. Khandpur is a research scientist, Department of Nutrition, University of Sao Paulo, Sao Paulo, Brazil, and a visiting scientist, Department of
Nutrition, Harvard TH Chan School of Public Health, Boston, MA.
Address correspondence to: Alyssa J. Moran, ScD, MPH, RD Department of Health Policy and Management, Johns Hopkins Bloomberg School of
Public Health, 624 N Broadway, Hampton House 363, Baltimore, MD 21205. E-mail: amoran10@jhu.edu
STATEMENT OF POTENTIAL CONFLICT OF INTEREST
No potential conflict of interest was reported by the authors.
FUNDING/SUPPORT
This research was supported by Healthy Eating Research, a national program of the Robert Wood Johnson Foundation.
ACKNOWLEDGEMENTS
The authors thank members of the Nutrition and Obesity Policy Research and Evaluation Network working group on Healthy Food Retail,
subcommittee on online retail (Lucia Leone, Diana Grigsby-Toussaint, Pasquale Rummo, Jared McGuirt, Alice Ammerman, Betsy Anderson
Steeves, and Stephanie Jilcott Pitts) for their contributions.

AUTHOR CONTRIBUTIONS
A. J. Moran conceptualized the research, obtained funding, developed instruments and protocols, analyzed data, and drafted the manuscript. G.
Headrick, C. Perez, and A. Greatsinger contributed to protocol development, collected and analyzed data, and provided critical manuscript
revisions. L. S. Taillie, L. Zatz, S. N. Bleich, and E. B. Rimm conceptualized the research, obtained funding, developed instruments and protocols,
and provided critical manuscript revisions. N. Khandpur conceptualized the research, obtained funding, developed instruments and protocols,
drafted the manuscript, and provided critical revisions.

2310 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12
RESEARCH

Table 5. Nutritional quality of the first fivea food and beverage items displayed in search results of 21 US online grocery retailers,
2019-2020

No. of Excessive Contains Excessive


products Poor Ultra-processed Excessive free other saturated
Search query recorded nutritional qualityb productc sodiumd sugarse sweetenersf fatsg

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
ƒ n (%)ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ
ƒ!
Bread 101 96 (95) 99 (98) 96 (97) 46 (46) 3 (3) 2 (2)
Canned fish 69 14 (20) 14 (20) 13 (93) 3 (21) 0 (0) 10 (71)
Entrée 97 94 (97) 95 (98) 93 (98) 23 (24) 0 (0) 57 (60)
Cheese 99 79 (80) 79 (80) 77 (97) 4 (5) 0 (0) 78 (99)
Chips 99 64 (65) 67 (68) 60 (90) 5 (7) 0 (0) 14 (21)
Canned corn 100 25 (25) 25 (25) 25 (100) 17 (68) 0 (0) 8 (32)
Salad dressing 95 92 (97) 92 (97) 91 (99) 38 (41) 1 (1) 52 (57)
Juice 95 36 (38) 42 (44) 3 (7) 36 (86) 1 (2) 0 (0)
Apples 91 2 (2) 2 (2) 0 (0) 2 (100) 0 (0) 1 (50)
Cookies 89 89 (100) 89 (100) 8 (9) 87 (98) 0 (0) 79 (89)
Pizza 99 97 (98) 99 (100) 97 (98) 5 (5) 0 (0) 87 (88)
Soda 101 99 (98) 100 (99) 16 (16) 80 (80) 19 (19) 0 (0)
Cereal 102 101 (99) 101 (99) 93 (92) 90 (89) 0 (0) 3 (3)
Oil 79 1 (1) 8 (10) 1 (13) 0 (0) 0 (0) 0 (0)
Milk 100 11 (11) 11 (11) 4 (36) 10 (91) 1 (9) 5 (45)
a
Researchers recorded up to five products; in some categories, fewer than five search results were displayed.
b
Products are defined as “poor nutritional quality” in the case that they are ultraprocessed and excessive in sodium, excessive in free sugars, containing other sweeteners, or excessive in
saturated fats.
c
Ultraprocessed products were defined using the NOVA classification system as products meeting the definition for category 4 (“ultra-processed”).24 These products were identified by
reviewing each product’s Nutrition Facts label and ingredients list.
d
Excessive sodium was defined as a ratio of sodium to energy 1. Value is % of ultraprocessed product.
e
Excessive free sugars was defined as energy from free sugar 10%. Value is % of ultraprocessed product.
f
Contains other sweeteners means the product contains any artificial or natural noncaloric sweeteners, as listed on the product’s ingredients list. Value is % of ultraprocessed product.
g
Excessive saturated fats was defined as energy from saturated fats 10%. Value is % of ultraprocessed product.

December 2022 Volume 122 Number 12 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS 2310.e1
RESEARCH

Food and beverage category Franckle and colleagues31 categories


Bread, Cereal, and Other Grains Bread; Cereal; Pasta, Rice, or Other Grain
Candy, Sweets, and Snacks Candy; Cold or Frozen Dessert; Sweet or Salty Snack; Sweet Bread, Cake, or Cookie
Condiments and Sauce Condiments, Sauces, and Salad Dressings
Fats and Oils Fat or OileSolid; Fat or OileLiquid
Fruit, Vegetables, and Legumes Fruit; Vegetable; Bean; Nut or Seed
Juice 100% Fruit Juice
Low-Calorie Drinks Lower-Calorie Beverage
Milk, Yogurt, and Cheese Milk or Milk Substitute; Yogurt; Cheese
Mixed Dishes and Prepared Foods Soup, Pizza, Convenience Foods; Convenience Foods (Hot)
Other Other: Food, Processed Soy/Meat Alternatives, Alcohol
Seafood, Meat, Poultry, and Eggs Red Meat, Processed Meat, Poultry, Seafood, Eggs/Egg Dishes
Sugary Drinks Sugar-sweetened Beverage
Unsweetened Drinks Unsweetened Beverage

Figure 2. Food and beverage categories used in content analysis of 21 US online food retail websites, 2019-2020

2310.e2 JOURNAL OF THE ACADEMY OF NUTRITION AND DIETETICS December 2022 Volume 122 Number 12

You might also like