**Background:** Online grocery shopping has nearly tripled in the US from $12.7 billion to $34.2 billion (2017–2021) and is projected to reach $47.6 billion by 2024–2026. As this channel grows, researchers need settings to evaluate nutrition interventions (e.g., warning labels, pricing changes) in online environments. Real-world retailers are often unwilling or unable to randomize consumers to different interventions, creating a need for naturalistic research platforms. This study evaluated a previously developed naturalistic online grocery store (Lola's Grocery) for feasibility, acceptability, and ecological validity.
**Methods:** From November 2021 to March 2022, 157 US adults were recruited via Craigslist; 144 completed both study visits (98% retention). Eligibility criteria included age ≥18 years, US residence, English literacy, online grocery shopping in the past 12 months, and internet access. Participants completed two online study visits (~30–60 min each) separated by ~1 week via Zoom. At each visit, they shopped in either the naturalistic online store (Lola's Grocery, offering ~23,000 products) or a real online store (Walmart.com), with store order randomized within participants. Participants used their own weekly shopping list and were instructed to select foods and beverages they typically buy. No payment was required for purchases. After shopping, participants completed survey measures on acceptability (e.g., satisfaction, ease of use) and realism. The naturalistic store automatically captured product data; for the real store, research staff recorded items from the cart. Spending was analyzed across 15 food and beverage categories (e.g., sugar-sweetened beverages, fruits and vegetables). Alcohol was excluded as it was unavailable in the naturalistic store. Analyses used Pearson's correlations and paired t-tests on proportional expenditures.
**Key Results:** The sample was diverse: mean age 39.5 years (SD = 11.0), 43% White, 38% Black or African American, 13% Latino(a) or Hispanic, 59% with household income <200% of the Federal Poverty Level. Acceptability was high: 100% of participants would participate in another study using the naturalistic store, 98% found most or all items on their shopping list, 83% found the store easy or very easy to use, 97% would pick similar products in the real world, 95% reported selections matched usual purchases, and 92% said the naturalistic store felt like a real online store. Mean total spending was $139.02 (SD = 83.96) in the naturalistic store vs. $156.04 (SD = 85.75) in the real store. Correlations in proportional spending between stores were moderate to strong across all 15 categories (range r = 0.36–0.67, all p < 0.001). The highest correlations were for meat and seafood (r = 0.67) and bread (r = 0.67), followed by dairy (r = 0.66) and fruits, vegetables, and legumes (r = 0.64). Few significant differences in mean proportional spending were found: participants spent slightly less on eggs (1.6% vs. 1.9%, p = 0.02) and salty snacks (4.6% vs. 6.3%, p = 0.001) in the naturalistic store, and slightly more on low-calorie drinks and juice (9.7% vs. 7.4%, p = 0.002). Purchases in the naturalistic store aligned with national data (e.g., 20.0% on fruits/vegetables vs. 19.8% in Consumer Expenditure Surveys; 21% vs. 21% on meat/seafood).
**Clinical Implications:** Naturalistic online grocery stores offer a feasible, acceptable, and ecologically valid platform for evaluating nutrition interventions in a controlled yet realistic setting. The high retention (98%), strong correlations with real-store behavior, and alignment with national purchasing data support the use of such platforms for randomized trials of policies like warning labels, pricing strategies, and product placement. Limitations include the sample being limited to prior online grocery shoppers and the use of hypothetical (non-purchased) selections. Future research should test whether delivering actual groceries from naturalistic stores is feasible.