**Background:** Live streaming platforms like Twitch combine real-time video content with interactive chatrooms, creating a unique environment for food marketing. Food advertisements on these platforms predominantly feature energy-dense, nutrient-poor products, and exposure to such marketing has been linked to poorer dietary behaviors and higher BMI in young people. Despite the prevalence of food marketing on Twitch, no prior research has examined how audiences interact with this marketing in real time. This study aimed to assess viewer engagement with a targeted food marketing campaign by the fast-food brand Wendy's, hypothesizing that the campaign would result in significantly more brand engagement, with more positive and fewer negative comments during and after the campaign period.
**Methods:** The study analyzed a 5-day Wendy's ad campaign called 'Never Stop Gaming' conducted on Twitch in December 2020, involving five major streamers. During sponsored periods, streamers displayed custom Wendy's logos, ran targeted video ads featuring themselves with custom-designed meals, and consumed these meals live on stream. Using the analytics platform Stream Hatchet, anonymized archived chat messages containing 'Wendy's' or 'Wendys' were collected from all Twitch chatrooms during the campaign period (December 8–12, 2020) and two comparison periods: one week before (December 1–5) and one week after (December 15–19). Messages were independently categorized by two research assistants as positive, negative, or neutral toward the Wendy's brand, with conflicts resolved by a third investigator (applying to 6.9% of messages). Chi-square tests were used to compare message proportions across time periods, with Bonferroni-corrected post-hoc analyses.
**Key Results:** Viewership of streams with 'Wendy's' in the title was dramatically higher during the campaign (1,433,645 hours watched) compared to before (257 hours) and after (542 hours). There were significantly more messages during the campaign period versus non-campaign periods (χ² = 28,688.71, P < 0.0001). Total messages were 52,015 during the campaign, compared to 16,286 before and 17,582 after. When considering all messages (positive, negative, neutral), the proportion differed significantly across time periods (χ² = 1,417.41, P < 0.001). During the campaign, neutral messages made up 71.3% of all messages (vs. 60.5% before and 61.0% after), positive messages 26.6% (vs. 33.3% before and 33.9% after), and negative messages 2.1% (vs. 6.2% before and 5.1% after). When analyzing only positive and negative messages, the proportion was also significantly different (χ² = 366.38, P < 0.001), with 92.5% positive during the campaign vs. 84.4% before and 86.9% after. Post-hoc analyses confirmed all pairwise differences were significant. Notably, the proportion of negative messages after the campaign (13.1%) was lower than before (15.6%), suggesting some sustained effect.
**Clinical Implications:** This study provides the first evidence of how interactive food marketing campaigns on live streaming platforms drive brand engagement in real time. The campaign not only increased overall brand mentions but shifted the sentiment ratio toward more positive and fewer negative comments, with effects persisting at least one week post-campaign. Given that increased engagement with food brands on social media has been associated with higher consumption of unhealthy foods and beverages, these findings raise concerns about the potential health impacts of such marketing, particularly on Twitch's predominantly young audience. The study highlights the need for future research on whether live streaming marketing acutely or longitudinally alters food choice and consumption, especially in children and adolescents. Policy implications include the difficulty of regulating digital marketing across borders and platforms, and the absence of age verification systems on most social media platforms, which limits the effectiveness of restrictions on marketing to minors.