Social media food marketing landscape of children living in Republic of Srpska, Bosnia and Herzegovina: Piloting the WHO CLICK framework, results and lessons learned

Abstract

Background

Monitoring food marketing that reaches children is key to effective implementation of any regulation. The Best-ReMaP European Union Joint Action aimed to test the feasibility of using monitoring protocols (developed by the World Health Organization’s Regional Office for Europe [WHO-Euro]) in the public health sector, to inform EU-wide monitoring plans. In Republic of Srpska, BiH, we piloted the WHO ‘CLICK’ Framework step, ‘L – Landscape of campaigns’, created to assess the nature of digital food and beverage marketing that children are potentially exposed to when using their preferred digital media.

Methods

This mixed methods pilot study analysed 235 Instagram and Facebook posts from 10 brands/products popular among children in Republic of Srpska, and 24 videos of two YouTube influencers, between January-April 2023. The WHO Nutrient Profile Model (1st edition, 2015) was used to assess ads’ eligibility to be marketed to children. With WHO-Euro monitoring Protocols we assessed the promotional techniques (ad’s creative power) used by brands and YouTube influencers. Finally, we assessed feasibility at every step of the process.

Results

Research capacity and coding were among the main identified challenges in the pilot. Our learnings contributed to updates of the WHO Protocols and the EU- framework guidance. Regarding the food marketing landscape, the most frequently advertised food categories on social media and in YouTube Influencers’ content were cakes, sweet biscuits and pastries (31% and 22.7% respectively) and chocolate and sugar confectionery (27.5% and 14.7% respectively). None of the social media ads were eligible to be marketed to children and only of 2 of the 24 YouTube influencers’ marketing instances were eligible. Most ads featured elements potentially appealing to teens, 88% on Facebook and 72% on Instagram. The most frequently used persuasive appeals on social media were reference to holidays, travel, or adventure (23%), followed by taste (12.5%) and premium/contest (11.5%). YouTube influencers predominantly presented branded foods (62.5%). Only one influencer video featured a marketing disclosure. Food cue contexts varied between the two influencers, depending on their target groups. One mostly posted videos filmed at home, with child appeals and referred positively to the food featured; the other presented food as a secondary object and in neutral rather than positive contexts.

Conclusion

The WHO-Euro ‘landscape’ Protocol is a feasible tool for Member States to evaluate digital food marketing. Facilitators were: capacity/resources, access to a knowledge/experience exchange, and regular protocol updates. The analysis shows that in Republic of Srpska, BiH, powerful marketing of unhealthy foods appealing to children takes place on children’s favoured social media platforms and via influencers with child-appealing content. Mandatory, government-led restrictions, accompanied by regular monitoring are needed to protect children’s health.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41043-025-01009-8.

Background

Evidence shows that almost all marketing to which children are exposed promotes food and non-alcoholic beverages (from now on, referred to as ‘food’) that are nutritionally poor and contain high levels of saturated fat, trans-fatty acid, sugar, and salt (HFSS). This contributes to unhealthy diets, childhood overweight and obesity and related non-communicable diseases (NCDs) and infringes children’s rights [1–3]. Increasing popularity of digital media and shifts in children’s media practices from television viewing to increased time spent online, including social media, content-sharing platforms (e.g., YouTube) and games (e.g., Fortnite) [4] has raised concerns among public health professionals and major health organization such as WHO and UNICEF about the influence of digital food marketing on children [4–6].

Food marketing through its extent and ‘power’ (creative strategies) [2] impacts children’s attitudes, preferences, purchase intention and consumption, adversely affecting their health [6]. The socio-ecological model provides a framework to understand how commercial actions in children’s macro-level environment, such as digital food marketing, shape their physical and social environments and impact diet-related outcomes, as well as their health and wellbeing [7]. At the physical level, digital food marketing reaches children in community locations (retailers, restaurants, outdoor advertising) and schools. At the social level, digital marketing enlists children’s peers and their online relationships with celebrities, influencers or brand and media characters [8]. To appeal to younger audiences, brands use powerful strategies such as celebrities, animated or brand equity characters which are popular among children [6]. Digital marketing heavily relies on digital influencers who often have a substantial fan basis including children, who find them more trustworthy than traditional celebrities [9]. This fast-changing field of digital advertising implies a need for constant monitoring of marketing activities in order to inform policies.

The most recently available survey of children’s digital media use in Bosnia and Herzegovina (BiH) was conducted in 2020; it revealed that 95% of children aged 7–18 years used YouTube and 85% also used at least one other social media or messaging platform, the most popular being the Viber messaging app (65%); Instagram (37%) was the main social media platform reported by older children 11–18 years while Viber and YouTube were commonly reported by younger children aged 7–10 (25% each). Almost all parents (96%) of children aged 11 to 18 years stated that their children had their own social media profile (Facebook, Viber, Instagram, and/or YouTube) [10].

International health organizations including WHO and UNICEF strongly recommend the implementation of restrictions on the marketing of unhealthy foods to children, to prevent obesity and diet-related non-communicable diseases [11, 12]. Evidence shows that comprehensive, mandatory, government-led restrictions of food marketing, covering all children until the age of 18, together with continuous monitoring are necessary to effectively protect children from their harms [13, 14]. In digital media to date, we are unaware of any such regulation having been implemented globally. However, in the UK, the implementation of the digital media restrictions in the Health and Social Care Act is expected in late 2025, which will limit advertising online of HFSS products to all users, although it has been noted that this regulation will not encompass brand marketing [15].

In BiH, guidelines on the regulation of the basic principles of audiovisual media services and radio media services are defined within the Code on Commercial Communication [16] from 2015 updated by the Communication Regulatory Agency of BiH in 2023. Article 7 “prohibits advertising that may jeopardize the health, psychological and/or moral development of minors” (Article 7, 1st paragraph). Other provisions of the Code state that commercial communications should not contain persuasive messages that can directly encourage children to buy advertised product/or facilitate pestering for the products, and that advertised HFSS food and beverages may not be shown during, or 15 min before or after, children’s programmes, and may not contain messages that highlight “positive qualities of nutritional aspects of such food and beverages” (Article 4, 6th paragraph). Other than this Code of Commercial Communication there is no current legal framework related to regulation or monitoring of marketing of unhealthy food in the digital landscape and to date, the nature of food and beverage marketing children in Republic of Srpska (RS) and BiH are exposed to has not been studied, including in digital marketing.

Monitoring the marketing of unhealthy products to children and adolescents is necessary to produce the evidence of children’s exposure, and is a fundamental element of any marketing restrictions to encourage compliance and allow enforcement. To strengthen monitoring and regulation, the World Health Organization (WHO) Regional Office for Europe has developed protocols for monitoring marketing of unhealthy products to children and adolescents for different media (TV and Internet). The CLICK Framework, introduced by WHO Europe in 2019, is a guideline to monitoring digital media specifically and it recommends that Member States Comprehend the digital ecosystem; develop insight into the Landscape of campaigns; Investigate exposure; Capture on screen; and engage in Knowledge sharing with communities, civil society and policymakers. The WHO Internet Social Media Brand Page Protocol and Internet YouTube Influencer Protocol used in this study are tools used to implement step L, landscape of campaigns. They assess the marketing campaigns of food brands/products popular among children deployed in media/settings frequently accessed by children [17, 18]. These methods study children’s potential exposure to food marketing. Although these methods or similar approaches to reporting on food marketing practices have been successfully used in research settings by specialists in the field [19, 20], the feasibility of monitoring by non-experts in general public health civil service settings has not been tested as far as we are aware.

The Best-ReMaP Joint Action (BRM-JA) was a three-year initiative (2020–2023) funded by the European Union’s (EU) Health Programme (2014–2020). Participating organizations (35 beneficiaries representing 24 European countries) collaborated on implementing pilot projects and generating practical lessons in the field of nutrition with a special focus on children and adolescents. BRM-JA sought to contribute to an improved quality of food supplied to EU citizens by adapting, replicating and implementing effective health interventions, based on best practices that have been proven to work in the areas of (1) food monitoring and reformulation; (2) framing of food marketing to children and adolescents and (3) public procurement of healthy food in public settings. One of the tasks of Work Package (WP) 6 was to test the feasibility of WHO food marketing monitoring protocols, to be implemented in an EU-wide monitoring programme. This required piloting and adapting available protocols to enable non-specialists to use them independently and deliver reliable outcomes [21].

In this context, the Public Health Institute, Republic of Srpska (PHI RS), piloted the protocols with a small sample of marketing content, implementing step L, Landscape of campaigns of the WHO CLICK Framework to monitor digital advertising of unhealthy food and beverage marketed to children.

Before this pilot, PHI RS were monitoring novices, with no experience regarding food marketing in general, or digital marketing in particular. In discussion with the BRM WP6 marketing leads, the Social Media Brand Page and YouTube Influencer Protocols were selected as most appropriate for piloting on a small sample to build capacity of the PHI RS team, collect initial digital marketing data, and test the feasibility of the protocols in a national public health setting.

Research questions:

• What is the nature of food and beverage promotion (i.e., types of marketed foods/beverage, their nutrient profiles, and creative marketing strategies used) on social media pages of brands popular among children and adolescents living in Republic of Srpska, BiH?

• What is the nature of food and beverage promotion by two You Tube influencers?

• How feasible are the WHO-Euro Protocols for novices to implement in a national public health setting, to monitor food marketing in digital media?

Method

Design

As a part of the Best-ReMaP JA piloting programme, the Public Health Institute Republic of Srpska (PHI RS) piloted the step “L – Landscape analysis” of the WHO CLICK monitoring framework using two WHO-Euro protocols: social media and YouTube influencer protocols [17, 22]. These provide guidance on assessing children’s potential exposure to food marketing in digital media, using exploratory mixed methods (quantifying the presence of marketing features and proportions of types of products/brands, and qualitative content analysis of creative strategies) to analyze HFSS food content placed in digital media by major food brands and YouTube influencers popular among children and adolescents living in Republic of Srpska, BiH. These WHO-Euro protocols have been applied in studies globally [19, 20]. Analysis includes types and nutritional profiles of marketed foods and creative marketing strategies (marketing ‘power’) used in branded content.

As the materials analyzed are all in the public domain on social media, and did not involve human participants, no ethical clearance was required.

Sample

Social media brand pages

Platform selection: Instagram and Facebook were selected during 2022 based on their popularity among children in BiH, according to the latest available data from a large survey conducted in 2020 [10].

Product/brand selection: As there was no publicly available official data on brands/products popularity and sales volumes in BiH, proxy data sources were combined to identify relevant brands to assess in social media. Data from ITC TRADE MAP - Trade statistics of international business development [23] were consulted to identify the food categories with high sales. Grey literature article searches were carried out for information on sales rates of food products in RS and BiH [24–26]. Digital media activity of brands such as number of followers/subscribers, post frequency, number of likes and comments were assessed on brand official web and social media pages in order to select brands/products with active social media platforms. Data were collected between January and April 2023.

For the ‘novice use’ pilot, we identified a sample of 235 posts from 10 social media pages (7 brands and 3 products) to monitor their marketing activity on their official Instagram and Facebook pages. We extracted a sample of the 10 most recent and/or pinned ‘unique’ posts from each page starting from the first post on the page [see Additional file 1 for details on selected pages]. Where identical posts were repeatedly posted in quick succession, these were excluded from content analysis to ensure capturing the full range of marketing strategies, but were included in brand activity rate calculations. Each post was screen-shot, assigned a unique number, and stored in the folder. Across the two social media platforms, we recorded 235 posts, 100 from Facebook and 135 from Instagram. Of these, 35 were repeat posts.

YouTube influencer marketing

To pilot novice use of the YouTube influencer protocol, we analyzed 24 videos, first selecting two influencers: Z&F, an influencer predominantly popular among children 12 years and younger (URL: https://www.youtube.com/@zagaifilip with 1.45 million subscribers according to data available in November 2022) and B.G. (URL: https://www.youtube.com/@bracog with 1.7 million subscribers according to data available in November 2022), an influencer popular among those aged 13–17 years; both located in RS, BiH [27, 28]. For each account we identified the most recent 12 videos featuring food items, starting with videos posted at the time of data collection (February 2023).

Coding – Social media brand pages and influencer marketing samples

WHO specifies exposure and ‘power’ as the two factors that together create the impact of food marketing [29]. We analyzed persuasive creative marketing strategies (‘power’) using the WHO-Euro social media and YouTube influencer protocols [22]. Variables coded for social media brand pages included the date content was posted, type and form of posts, food products featured (if any), nutritional information, eligibility to be marketed to children (according to the WHO Nutrient Profile Model), and persuasive strategies used to appeal to consumers. The ‘primary’ and ‘secondary’ persuasive appeals (defined as the two main themes or topics of the advertisement) were selected for each post from eighteen categories (including taste, fun and holiday/travel/adventure). Other ‘power’ variables included the presence of brand logos, products and packaging, use of brand equity or other characters, links to events or special days, premium offers/sponsorships, depiction of physical activity and nutritional claims. The presence of elements potentially appealing to children (12 and under) and teens (13–17 years) as defined by the protocol, was also coded [22].

Similar information was coded for YouTube influencer but also included the food/beverage cue start and end time in the video. Power variables analyzed included cue brand status, cue context (settings where influencer present food), cue description (positive, negative/neutral adjectives or tone) and cue presentation (consumption/(non)verbal presentation) etc. A list of all variables coded can be found in supplementary materials [see Additional file 2].

The marketing posts and videos were classified as healthy (and therefore eligible to be marketed to children) or unhealthy (and not eligible) using the WHO-Euro Nutrient Profile Model (NPM) v.1 [30] as v.2 of the WHO had not yet been published at the start of the study. Where more than one product was featured in the post/video, all products were coded and if at least one product was classed as “unhealthy/not permitted”, the advertisement was categorized as ‘not permitted’. For brand marketing, where no specific product was present, we visited the brand’s official national (or, if unavailable, international) website and identified the most dominant product on the home page; this was assigned to the post. To nutrient profile products, we used nutrition factsheets on official websites (preferably local; international if not available).

Inter-rater reliability testing

All posts were coded independently by two researchers. Together with a third independent researcher, they discussed the codebooks and definitions of each variable. The first and fifth authors coded 10% of posts simultaneously (10 each from FB and Instagram). Interrater reliability (Kappa) was calculated for an initial set of 13 categorical variables1 considered prone to more subjective assessment. The average Kappa for this initial set was 0.69 implying substantial agreement, but such that the individual values were investigated to identify areas of lower agreement. The highest Kappa values were present among variables related to brand ad, food category code, post type and does the ad appeal to children (12 and under) (all Kappa = 1, p < 0.001), other character (Kappa-0,835, p < 0.001) image of product (0.783, p < 0.001) and celebrity endorsers (Kappa-0.634, p < 0.01). The lowest values were found at first for primary (0.441, p < 0.001) and secondary (0.412, p < 0,001) persuasive appeals; coders carefully discussed discrepancies and after another coding the final Kappas were higher for primary (0.848, p < 0.01) and secondary (0.667, p < 0.01) appeals.

Intercoder reliability for YouTube influencers was calculated on 5 videos (3 from Z.F., 2 from B.G.). Interrater reliability (Kappa) was calculated for 6 variables (nutrient profile status, cue brand status, cue context, cue description and cue presentation), for which the average Kappa was 0.85 implying good agreement. The lowest agreement was for cue context (Kappa- 0.722, p < 0.01) and cue description (Kappa-0.545, p = 0.171). Cue description was found to be particularly demanding to code due to influencers’ ambivalent messages, simultaneously positive and negative, toward food items presented in the video.

After the inter-rater reliability stage, throughout the data analysis, coders constantly discussed coding with each other and consulted with BRM-JA WP6 experts (third and fourth authors) who were authors of the WHO-Euro Protocols.

Data analysis

All variables were coded in the WHO-Euro social media brand page and YouTube influencer video templates (Microsoft Excel). Descriptive analyses and inferential tests (Chi-square, Fisher’s exact tests) were performed using SPSS (Statistical Package for Social Science) version 16, to answer research questions, including any significant differences in marketing appeals between platforms and between age groups.

Results

Social media brand pages (Instagram and Facebook)

We extracted 235 advertising posts (100 Facebook, 135 Instagram) of which 200 were unique and represented the final content analysis sample; of these, 44 (22%) were brand ads that promoted the brand but did not specifically feature any product.

The categories marketed most frequently were cakes, sweet biscuits and pastries (31%) and chocolate and sugar confectionary (27.5%) (see Table 1 for frequencies).

Table 1.

Foods and beverages present in the brand page posts categorized by WHO-Euro NPM

	Social Media Platform		Total (n = 200) n (%)
	Facebook (n = 100) n (%)	Instagram (n = 100) n (%)
Chocolate and sugar confectionery, energy bars, sweet toppings and desserts	30 (30%)	25 (25%)	55 (27.5%)
Cakes, sweet biscuits and pastries; other sweet bakery wares, and dry mixes for making such	28 (28%)	34(34%)	62 (31%)
Savoury snacks	10 (10%)	10 (10%)	20 (10%)
Beverages - Juices (Includes 100% fruit and vegetable juices; juices reconstituted from concentrate, and smoothies)	1(1%)	1(1%)	2 (1%)
Beverages - Energy drinks	10(10%)	10 (10%)	20 (10%)
Beverages - Other (Includes cola, lemonade, orangeade; sweetened beverages, mineral and/or flavoured waters (including aerated) with added sugars or sweetener	10 (10%)	10(10%)	20(10%)
Edible ices (Includes ice cream, frozen yoghurt, iced lollies and sorbets)	10 (10%)	10(10%)	20 (10%)
Breakfast cereals (Includes oatmeal; cornflakes; chocolate breakfast cereals; mueslis)	1(1%)	0(0%)	1(0.5%)

None of the marketing posts would be permitted for advertising to children (according to WHO-Euro NPM).

Most food/beverage advertising posts were found to have elements appealing to teens, significantly more so on Facebook (88%) than Instagram (72%) (χ²=8, df = 1, p < 0.05). Elements appealing to children were found in almost half of posts, at similar rates on Facebook (50%) as Instagram (44%) (χ²=0.723, df = 1, p = 0.395) (Fig. 1).

Fig. 1.

Social media brand pages posts appealing to children and teens

The use of promotional techniques

On both social platforms holiday, travel or adventure was most frequent primary persuasive appeal (23%), followed by taste (12.5%) and premium/contest (11.5%). Premium/contest and taste were more frequently identified on Instagram (16% and 15% respectively) than Facebook (7% and 10% respectively). Some appeals were more frequently present on Facebook than Instagram: links to events or entertainment (11% vs. 3% on Instagram) and holiday, travel or adventure (Facebook: 25%; Instagram: 21%) but only the last difference was statistically significant (χ²=42.745, df = 20, p < 0.01).

Some appeals (not in Fig. 2) featured in fewer than 5% of posts: quantity, health/nutrition, energy, unique, family relationships, general superiority, peer status, friendship, offers choices/options, new product introduced, humor, magic/fantasy, novel or surprising feature.

Fig. 2.

Frequency of primary persuasive appeals used in social media posts on Facebook and Instagram

We found the classification of primary and secondary appeals was challenging, so to identify the most common appeals, we combined the frequencies for these and created a binary for each ‘power’ variable (appeal present/absent). In this measure, a post could therefore have more than one appeal. Using this method, we identified that holiday/travel; fun; and taste were the three most common appeals, and that reference to convenience was significantly more present in Instagram (16%) than Facebook posts (7%) (χ²=3.979, df = 4, p = 0.046). (See Table 2 for the top 10 appeals).

Table 2.

Most common identified appeals on social media platforms

Main appeals identified	Social Media Platform		Total1 (n = 200*) n (%)
	Instagram (n = 100) n (%)	Facebook (n = 100) n (%)		p 2
Holiday/ travel	23 (23%)	33 (33%)	56 (28%)	0.115
Taste	21 (21%)	19 (19%)	40 (20%)	0.724
Fun	20 (20%)	16 (16%)	36 (18%)	0.462
Premium or contest	19 (19%)	10(10%)	29(14.5%)	0.071
Enjoyment	17(17%)	12(12%)	29(14.5%)	0.315
Convenience	16(16%)	7(7%)	23 (11.5%)	0.046*
Romance or sex appeal	11(11%)	8(8%)	19 (9.5%)	0.469
Link to event or entertainment	5(5%)	12(12%)	17(8.5%)	0.076
Friendship	5 (5%)	9 (9%)	14 (7%)	0.268

^*please note, that in this outcome, appeal could be classed as either in primary OR secondary, therefore the sum of individual s is higher than total number of posts

¹Percentages (%) presented in columns correspond to category of answer with identified appeal noted in each row. ²p values for Chi-square with values < 0.05 were considered significant (*)

Brand logos were shown in 58.5% of posts, significantly more frequently on Instagram (80%), than Facebook (37%) (χ²=38.081, df = 1, p = 0.000). Celebrities (mainly entertainment, n = 8 and sports celebrities, n = 8) were present in 18% of posts. Other celebrities (n = 16) were fashion/ lifestyle bloggers and athletes. Licensed characters were only found on Facebook posts (7%). Nearly a third (30.5%) of posts were linked to “special days” through the year, revealing the seasonal/thematic character of advertising across time (New Year, International Women’s Day, Valentine’s day, etc.), and significantly more on Facebook (38%) than Instagram (23%) (χ²=5.307, df = 1, p = 0.021). Premium offers with competitions or contests were more prevalent on Instagram (20%) than Facebook (10%) posts (χ²=3.922, df = 1, p = 0.048). Physical activity was present in 10% (n = 20) of posts, mostly depicting high-energy/excitement sports activities sponsored by one energy drink brand. Finally, health claims were identified only on two Facebook posts, one related to natural ingredient and other as provider of essential nutrients.

See Table 3 for marketing techniques identified for advertising posts across both platforms.

Table 3.

Marketing techniques used in social media posts

	Instagram n = 100	Facebook n = 100	Total* n = 200	p 2
	n (%)	n (%)	n (%)
Identifying
Brand logo shown (with post itself)	80 (80%)	37 (37%)	117 (58.5%)	< 0.001*
Post include an image of the packaging	77 (77%)	79 (79%)	156 (78%)	0.733
Image of product included in the post (not just the packaging)	40 (40%)	35 (35%)	75 (37.5%)	0.465
Characters used
Celebrity present on the post	15 (15%)	21 (21%)	36 (18%)	0.269
Brand equity character present	3 (3%)	9 (9%)	12 (6%)	0.134
Licensed character present	0 (0%)	7 (7%)	7 (3.5%)	0.014*
Linking
Post linked to “special day”	23 (23%)	38 (38%)	61 (30.5%)	0.021*
Post linked to entertainment event	10 (10%)	8 (8%)	18 (9%)	0.621
Post linked to sport event	7 (7%)	9 (9%)	16 (8%)	0.602
Premium offers/ sponsorships
Premium offers with competition or contest	20 (20%)	10 (10%)	30 (15%)	0.048*
Sponsorship mentioned in the post	10 (10%)	4 (4%)	14(7%)	0.164
Presence of physical activity
Physical activity depicted	8 (8%)	12 (12%)	20 (20%)	0.346

^*Percentages (%) presented in columns correspond to category of answer with present marketing technique noted in each row. ²p values for Chi-square and Fischer Exact Test with values < 0.05 were considered significant (*)

Power of marketing identified as appealing to children and teens

In ads identified as appealing to children, the most frequent appeals were holiday/travel (over a quarter of ads) and taste and premium/contest (each in about a fifth of ads) (Table 4). The most frequent appeals in ads appealing to teens were also holiday/travel, in over a quarter of ads, followed by fun and by taste (a fifth of ads).

Table 4.

Most common identified appeals among children and teens

Main appeals identified	Appealing to children (12 and under) n (%)		Appealing to teens (13–17 y) n (%)		Total n(%)
	no	yes	no	yes
Holiday and travel	30(28.3)	26(27.7)	12(30)	44(27.5)	56(28)
Taste	19(17.9)	21(22.3)	11(27.5)	29(18.1)	40(20)
Premium or contest	10(9.4)	19(20.2)	2(5)	27(16.9)	29(14.5)
Fun	24(22.6)	12(12.8)	2(5)	34(21.2)	36(18)
Link to event or entertainment	10(9.4)	7(7.4)	2(5)	15(9.4)	17(8.5)
Romance or sex-appeal	18(17)	1(1.1)	7(17.5)	12(7.5)	19(9.5)
Convenience	15(14.2)	8(8.5)	10(25)	13(8.1)	23(11.5)
Enjoyment and satisfaction	19(17.9)	10(10.6)	12(30)	17(10.6)	29(14.5)
Friendship	8(7.5)	6(6.4)	0	14(8.7)	14(7)

^*Percentages (%) presented in columns correspond to category of answer with present appeals noted in each row for posts appealing for children and teens

We also examined the ten most frequently used power strategies (as identified above) in posts identified as appealing to children and in those appealing to teens. Posts featuring contests (20.2%) were significantly more likely to feature in ads appealing to children than those with no child appeal (9.4%) (χ²=4.669, df = 1, p = 0.031). Romance or sex-appeal was significantly less frequent among posts appealing to children (1.1%) than posts not appealing to children (17%) (χ²=14.682, df = 1, p < 0.001). Ads with teen appeal significantly more frequently featured appeals such as fun (χ²=5.725, df = 1, p = 0.017), enjoyment (χ²=9.689, df = 1, p = 0.002) and convenience (χ²=8.954, df = 1, p = 0.003) compared to those without teen appeal (Table 4).

Next, we analyzed and compared the frequency of brand ads in posts classified as appealing to children and to teens (Fig. 3). We found that brand marketing was significantly less frequent in ads appealing to children 12 or under compared to those without child appeal (8.5% vs. 34%) (χ²=18.807, df = 1, p = 0.000). The same was not observed for posts appealing to teens (21.9% vs. 22.5%) (χ²=0.007, df = 1, p = 0.932).

Fig. 3.

Percentage of brand ads in posts classified as appealing to children and teens

Results of you tube social media influencers

Twenty-four videos with food items, 12 from each of the two YouTube influencers, were extracted for analysis. Z&F are young parents and they create shorter videos (average duration- 7min56s) presenting funny products and creative games for children branded with their names and visuals (e.g. croissant filled with chocolate cream called Z&F); B.G. creates longer videos (average − 28min16s), depicting funny situations and everyday adventures with companions. In sum, 450 min of video content was analyzed for two influencers.

The influencers Z&F, appealing to younger age groups, featured food items more frequently than B.G. did. The frequency of posting videos with food items/products per influencer is shown in Table 5.

Table 5.

YouTube influencer videos and frequency with food items

YouTube influencers	Total (no. of analyzed videos)	Number of days between first and last analyzed video	Frequency of videos with food items (every x days)
Z&F	12	123	10.25
B.G.	12	545	45.42

The range of food products/items featured per video was 1–11. These were analyzed to identify categories and nutrient profiles, according to WHO-Euro NPM. Most items were cakes, sweet biscuits, and pastries (22.67%), chocolate and sugar confectionary (14.67%), and beverages with added sugar and sweeteners (14.67%); see Table 6.

Table 6.

Foods and beverages present in the videos categorized by WHO NP food category code

WHO NP Food category code	Z&F	B.G.	Total
	n	n	n (%)
Chocolate and sugar confectionery, energy bars, sweet toppings and desserts	5	6	11 (14.67%)
Cakes, sweet biscuits and pastries; other sweet bakery wares, and dry mixes for making such	8	9	17 (22.67%)
Savoury snacks	2	4	6 (8%)
Beverages - Juices (Includes 100% fruit and vegetable juices; juices reconstituted from concentrate, and smoothies)	1	1	2 (2.67%)
Beverages - Milk drinks	1	2	3 (4%)
Beverages - Energy drinks	0	1	1 (1.33%)
Beverages - Other (Includes cola, lemonade, orangeade; sweetened beverages, mineral and/or flavoured waters (including aerated) with added sugars or sweetener	0	11	11 (14.67%)
Edible ices (Includes ice cream, frozen yoghurt, iced lollies and sorbets)	1	0	1 (1.33%)
Yoghurts, sour milk, cream and other similar foods	0	1	1 (1.33%)
Cheese; grated or powdered cheese; cottage cheese; processed cheese spreads)	1	0	1 (1.33%)
Ready-made and convenience foods and composite dishes	2	2	4 (5.33%)
Butter and other fats and oils (Includes butter; vegetable oils; margarines and spreads	0	1	1 (1.33%)
Bread, bread products and crisp breads	0	1	1 (1.33%)
Fresh and frozen meat, poultry, fish and similar	0	3	3 (4%)
Processed meat, poultry, fish and similar	0	2	2 (2.67%)
Fresh and frozen fruit, vegetables or legumes	3	2	5 (6.67%)
Processed fruit, vegetables and legumes	1	2	3 (4%)
Sauces, dips and dressings	0	2	2 (2.67%)
Total	25	50	75 (100%)

Table 6. Foods and beverages present in the videos categorized by WHO NP Food category code.

Of the 24 videos featuring foods/beverages, 22 (91.7%) featured items not permitted for advertising to children according to the WHO NPM, with the exception of two videos (8.3%) where Z&F briefly presented fresh fruit and vegetables. A further 9 (40.9%) of these 22 videos also featured healthy foods, but according to the WHO NPM the presence of at least one non-eligible product disqualifies the ad from being marketed to children. Furthermore, videos of B.G. featured alcoholic drinks such as beer and wine (n = 3, 25%). Over half the videos (n = 14, 58.3%) also featured unbranded foods whose nutrient profile is hard to determine (slice of white bread, pizza, spices, etc.).

The marketing strategies and approaches varied in the influencer videos (Table 7). Although nearly two-thirds of the videos presented branded food products (62.5%), their cue contexts differed. Z&F mostly posted videos filmed at home (n = 9, 75%) with bright colors and vivid pictures appealing to children and referred positively to the food featured (n = 8, 66.7%). In contrast, B.G. presented food as a secondary object and in neutral (n = 6, 50%) rather than positive (n = 4, 33.3%) contexts including alcoholic drinks which were also presented neutrally. In two videos candies and beverages with non-sugar sweeteners were described as “healthy”. In another two in the local supermarket, food was present in the background and B.G didn’t consume or verbally refer to it.

Table 7.

Food cues in videos and marketing techniques used by influencers

	Z&F n = 12	B.G. n = 12	Total n = 24
	n (%)	n (%)	n (%)
Food cue brand status
Branded	6 (50%)	9 (75%)	15 (62.5%)
Food retail establishment	0	1 (8.3%)	1 (4.2%)
Unbranded	4 (33.3%)	1 (8.3%)	5 (20.8%)
Miscellaneous	2 (16.7%)	1 (8.3%)	3 (12.5%)
Food cue context
Supermarket	1 (8.3%)	2 (16.7%)	3 (12.5%)
Home	9 (75%)	3 (25%)	12 (50%)
Other	2 (16.7%)	7 (58.3%)	9 (37.5%)
Food cue description
Positive	8 (66.7%)	4 (33.3%)	12 (50%)
Negative	0	2 (16.7%)	2 (8.3%)
Neutral	4 (33.3%)	6 (50%)	10 (41.7%)
Food cue presentation
Consumed and verbally referenced	5 (41.7%)	5 (41.7%)	10 (41.7%)
Consumed and not verbally referenced	2 (16.7%)	1 (8.3%)	3 (12.5%)
Not consumed and verbally referenced	5 (41.7%)	4 (33.3%)	9 (37.5%)
Not consumed and not verbally referenced	0	2 (16.7%)	2 (8.3%)
Why food cue featured
Non-explicit marketing	12 (100%)	11 (91.7%)	23 (95.8%)
Paid endorsement	0	1 (8.3%)	1 (4.2%)

In all the videos, paid endorsement was only disclosed by one influencer (B.G), for flavored wafers (4.2%).

Feasibility of protocol use

Finally, we evaluated the feasibility of using the WHO protocols for novice users. Overall, the WHO-Euro Internet Social Media Brand Page and YouTube Influencer Protocols [22] facilitated monitoring activities, by providing a structured, standardized and comprehensive guidance, accompanied by coding sheets, and set the opportunity to collect data related to food marketing in Republic of Srpska for the first time. The EU Best ReMap Joint Action WP 6 team provided guidance and expertise to the novice PHI RS food marketing monitors. However, certain challenges were also encountered.

During the pilot’s preparatory stage, PHI RS encountered challenges in finding recent, reliable sources of information on brands/products and influencers popular among children and adolescents in the Republic of Srpska; no official data or studies were identified as the protocols note is often the case. The best available grey literature sources were used instead, as recommended by protocols. To inform further monitoring and policy development across the EU, per-country focus groups and surveys of representative sample of children and parents are needed to explore children’s media and eating habits (e.g., following step C ‘comprehend ecosystem’ of the WHO CLICK Framework).

A significant challenge related to available resources (funding and researcher time) resulted in the small sample in this pilot. Food marketing monitoring studies require a significant amount of manual coding; this is time consuming and requires some expertise in nutrition, food marketing, digital media and mixed methods analyses. In the pilot, the availability of experts and knowledge exchange with other piloting teams facilitated the process. In the future, full scale studies will require adequate resources and expertise in the research team to examine larger samples.

Despite significant detail in the protocols, the piloting revealed elements requiring further clarification or amendment. Going beyond exposure to assess the ‘power’ of marketing (as recommended by WHO 2012) [29] including coding for persuasive appeals is a qualitative process. This required regular discussion and verification between coders and with experts for the coding process to achieve acceptable levels of agreement. Particularly challenging was identifying the primary and secondary persuasive appeals of the ad. It became apparent, that although the presence of an appeal was usually not difficult to assess, as posts frequently featured multiple appeals, specifying the dominance of one over another was challenging. After discussions between coders and with the WP6 team we reached an agreement and amendments were made to the protocol where most of the appeals have now become binary (presence/absence), without the indication of primacy. This amendment has now been made to the WHO-Euro/EU Protocols.

The lack of availability of information on the presence of some nutrients required to profile the ads as eligible or not to be marketed to children such as added sugars and non-sugar sweeteners presented a challenge as in Republic of Sprska mentioned nutrients are not indicated on product nutritional sheet and resulted in missing eligibility data for some products.

In terms of format, video content was more demanding and time consuming to analyse than still images, as multiple different food products featured (up to 11 in one post) including unbranded items where nutrient profiles were hard to determine. Additionally, defining the cue description was difficult when different food products/items were simultaneously present in positive and negative contexts; and not all items featured in WHO-Euro NPM food categories.

Overall, however, despite the above challenges outlined, we found the WHO-Euro monitoring protocols feasible for use by non-experts in a general civil service public health setting, due to the clear, structured guidance provided.

Discussion

This was the first study of food and beverage marketing on social media sites popular with children in Republic of Srpska. The study successfully applied the WHO-Euro/EU food marketing monitoring protocols showing their feasibility for use by novice food marketing monitors in public health settings, with minor amendments and expert support. We found that in the Republic of Srpska, brands and products popular with children engage in highly impactful marketing on digital media platforms popular with children (social media and YouTube) and consists of mostly unhealthy products. Almost all posts/videos by YouTube influencers featuring food contained unhealthy products, yet only one came with the disclosure indicating it is a paid promotion.

All posts advertised products that should not be permitted to be marketed to children, most frequently cakes, sweet biscuits and pastries (31%) and chocolate and sugar confectionary (27.5%). This is in line with other studies conducted by Culp et al., Kent et al., Tan et al. that typically show a strong dominance of HFSS products [31, 32, 33] and scarcity of healthy food promotion [34, 35]. This is not surprising, considering that most (nearly 90%) of sales of major food brands come from unhealthy products [36]. Different food and beverage products from neighborhood countries are dominantly present in Bosnia and Herzegovina market [37] together with social media sites used for their promotion, so analyzed food marketing reveal potentially same marketing strategies used among brand/products that share market among countries in the Balkan’s region.

Brands used a set of powerful marketing methods, the most frequent appeals being holiday/travel, taste, contest or a premium item giveaway – and fun for ads appealing to teens specifically. Elsewhere in other countries the most frequently used marketing techniques are similarly related to competition or humour [38], fun, enjoyment/ satisfaction or taste [39] implying that marketing strategies have similarities but may potentially also vary in some respects between different socio-cultural contexts. Approximately one third (30.5%) of the posts were linked to “special days” through the year, revealing the seasonal/thematic character of advertising (New Year, International Women’s Day, Valentine’s Day etc.) and popular global activities (FIFA Men’s World Cup); this type of marketing strategy was more present on Facebook (38%) than Instagram (23%). This shows that this type of seasonal and event related marketing is common, frequently used technique by brands (indeed, recommendations in earlier protocols and food marketing studies to avoid special seasons and events in monitoring might lead to a significant under-estimation of the extent and power of marketing children are exposed to). Such marketing is also very powerful, as special days and seasons are associated with strong, positive emotions, time spent with family and friends, national identity, and sports such that these elements become associated with unhealthy food and brands in consumers’ perception.

As different social media platforms offer different experiences to users, monitoring studies reveal different advertising strategies between Facebook and Instagram. Engagement strategies are more prevalent on social platforms that target younger audience [19], which we found with Instagram featuring more prevalent promotional techniques (premium offers with competitions or contests). In comparison to Facebook, visual strategies such as brand logo were more present on Instagram, facilitating consumer awareness on brands products, consumption and loyalty during life course [40].

Social media posts regularly feature celebrities based on the assumption that audiences, wishing to emulate celebrity lifestyles [41], form a preference for celebrity-endorsed brands [42]. In our study, celebrities featured in nearly fifth of all posts. Studies show that not only the presence of celebrities in marketing content increases children’s intake of marketed food, but even just exposure to a celebrity known to promote the product, outside of the food context, triggers consumption [43, 44].

Marketing campaigns we identified in social media in Republic of Srpska often featured elements likely to appeal more to teens (80%) rather than children (47%), with marketing on Facebook mostly seeming to target teens. Similar findings can be seen in a recent UNICEF study of unhealthy digital food marketing to children in the Philippines [39] where digital marketing posts appealed more to older children, compared to younger children. This is likely the response to the target audience, since Instagram and Facebook are more popular among teenagers, and YouTube among younger children in Republic of Srpska [10].

Similarly, our study indicates that food content featured by influencers popular among children and teens in Republic of Srpska is largely unhealthy, with most being cakes, sweet biscuits and pastries (22.67%), chocolate and sugar confectionary (14.67%) and beverages with added sugar and sweeteners (14.67%). You Tube influencers predominantly present branded food products (62.5%) potentially impacting children’s food brand preferences as well as their short-term consumption [45]. This reflects the absence of clear regulation in field of digital marketing in Republic of Srpska.

One of the influencer channels focused on younger children, mostly posting videos filmed in the influencers’ own home. They almost always referred positively to the foods featured, in videos full of bright colours and vivid pictures appealing to children. Influencers depicting their everyday home activities which appear as authentic have long been common in social media [46] and have been shown to have strong persuasive effect [47]. Influencers’ advertising messages are often perceived as authentic messages, that go beyond promoting products and services to influence their followers’ health-related attitudes and habits [48, 49]. In contrast, the second influencer, targeting teens, depicted food products/items in non-home contexts, and food was not the main focus. In two videos, candies and beverage with non-sugar sweeteners were described as “healthy”. As a feasibility pilot, a comprehensive account of the full extent and nature of digital influencer food marketing in Republic of Srpska was not in this study’s remit. However, we note that previous studies have confined themselves to two influencers (see e.g., Coates et al., 2019) and the total sample for the influencers comprised 450 min of video content.

From a feasibility perspective, we note several limitations related to the preparation process and selection of brands/products of food and beverage for monitoring. The WHO-Euro Internet protocol provided a clear description of preparation steps, but as it notes is often the case, local data related to brands/products popularity or sales were not available. We therefore instead, to identify brand/product popularity, analyzed official websites, identified numbers of followers and advertising dynamics on official social media pages, and newspaper articles. The WHO-Euro Protocols recognize that information available to researchers will vary with each context, and allow for such variations in the process, providing detailed guidance for monitoring activities, with introductory and theoretical framework, decision making threes, Excel files with templates for coding among many potential sources of marketing to children [22]. Such a framework provides a clear platform for conducting rigorous public health research needed to recognize and understand target marketing as a specific contextual influence on the health of children and adolescents that will support identification and implementation of corrective policies.

The WHO-Euro Protocols also recommend surveying children to determine the most popular media channels for monitoring, adding the possibility of running focus groups with children and young people to better explore media and marketing ecosystems [21] which can be considered in designing further studies in this fast-changing field. Again, this was beyond the scope of the present study but would be a valuable activity in the future. The scope of monitoring activities often depends on available resources such as time, finance, monitoring skills and training [21] but, resources depending. Future studies should increase the sample size of brands/products and influencers, and monitor marketing activities across different channels (TV, outdoors and community venues, cinema, etc.) and other digital media to provide representative and comprehensive results.

Conclusion

This pilot, and first study exploring marketing activity of selected brand/products in Republic of Srpska, BiH, has provided an indicative baseline and has built capacity for designing and conducting similar studies with bigger samples, and/or of actual exposure using children’s screen recordings and/or monitoring other channels of communication. It also provided a proof-of-concept for the application of the WHO-Euro/EU Food marketing monitoring protocols in local, non-research settings. This project, along with other studies piloting the WHO CLICK Framework, informed updates of the existing protocols and tools, and creation of new ones. These tools now, adapted for use in non-academic, civil-services public health settings, can be used to provide evidence for policy design, enforcement and evaluation. Considering the dynamic character of the digital media landscape, including marketing, regular validation and updates should take place to ensure their feasibility and effectiveness.

Following this piloting exercise, especially in the light of identified lack of recent data on media use, next steps should include a full-scale monitoring activity, that along with current findings, will inform development of policies and strengthen advocacy efforts in the region. A regulatory taskforce should be established and multi-stakeholder meetings should take place to accelerate policy development.

The food marketing landscape we identified highlights an urgent need to review marketing policies in the country and urgent action to protect children from its harms by supporting development and straightening regulations that control children’s exposure and protect children of potential harmful food marketing. Following the new WHO Guideline on Policies to protect children from the harmful impact of food marketing [50] such restrictions should be government-led and mandatory, protect children of all ages up to 18 years, be comprehensive in scope (including influencer marketing) and protect children from all marketing they are exposed to, not just marketing targeted at them. Independent and regular monitoring must aid any restrictions to allow for enforcement. Development of user-friendly materials about the research findings through partnership with parents, young people, policy–makers and civil society can create a platform for advocating for change, raising awareness and influencing policy. Influencer food marketing and the transparency of their financial interests should be included in any planned restrictions given the role of this new form of marketing a common and powerful means of marketing to children and teens.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

HFSS

High Fat Salt and Sugar

NCDs

Non-communicable diseases

BiH

Bosnia and Herzegovina

WHO

World Health Organization

UNICEF

United Nations International Children’s Emergency Fund

CRA

Communications Regulatory Agency

RS

Republic of Srpska

CLICK

Comprehend the digital ecosystem, Landscape of campaigns, Investigate exposure, Capture on-screen, Knowledge sharing

BRM-JA

Best-ReMaP Joint Action

PHI RS

Public Health Institute Republic of Srpska

ITC

International Trade Centre

NPM

Nutrient Profile Model

SPSS

Statistical Package for Social Science

WP

Working Package

EU

European Union

Author contributions

J.N. D.S. and M.M and M.T.G contributed to the writing of the article. M.M, M.T.G. and J.N. designed the study; J.N. and D.M. collected and analyzed the data; J.N., D.S and D.M. contributed to the writing study report. All authors have read and approved the final manuscript.

Funding

This paper draws on the authors’ joint contribution to activities of Working Package (WP) 6 of the Best-ReMaP project funded by the European Union’s (EU) Health Programme (2014–2020) and participating organizations.

Data availability

The datasets analyzed in the study are available from the corresponding author.

Declarations

Ethics declaration

Ethics declaration: not applicable.

Consent to publish

Consent to Publish declaration: not applicable.

Consent to participate

Materials analyzed in the paper are all in the public domain on social media, and did not involve human participants.

Competing interests

The authors declare no competing interests.