Selected Papers of #AoIR2024: The 25th Annual Conference of the Association of Internet Researchers Sheffield, UK / 30 Oct - 2 Nov 2024 Suggested Citation (APA): Walton, M. & Aderibigbe, D. (2024, October). Hairy industries The politics of advertising hair products and services to South Africans on Facebook and Instagram. Paper presented at AoIR2024: The 25th Annual Conference of the Association of Internet Researchers. Sheffield, UK: AoIR. Retrieved from http://spir.aoir.org. HAIRY INDUSTRIES: THE POLITICS OF ADVERTISING HAIR PRODUCTS AND SERVICES TO SOUTH AFRICA ON FACEBOOK AND INSTAGRAM Marion Walton marion.walton@uct.ac.za Deborah Aderibigbe ADRIRE001@myuct.ac.za Introduction In the white supremacist discourses of apartheid South Africa (SA), hair was a key signifier of race. Local media industries still encourage Black African readers to “aspire to whiteness as a social and aesthetic category” (Pilane, & Iqani, 2022) and to idealise long, straight hair in particular (Madlela, 2018), all the while promoting products that alter hair dramatically and possibly harmfully (Spector et al., 2023). Globally, Black anti- racist activists have challenged these hegemonies (Tate, 2007; Sobande, 2020) including in South Africa, where stores were forced to close and Unilever’s TRESemmé products were removed from shelves after an outcry about a racist advertising campaign (“TRESemmé,” 2020). Women’s hair routines can involve considerable expense and aesthetic labour (Elias et al., 2017), and this is particularly so for many Women of Colour (WoC). Nonetheless, informal skill development and formal work in hair and beauty industries are also vital to many women’s livelihoods and self-expression. This paper shows how minor brands, small businesses and micro-entrepreneurs in SA use targeted ads on Meta platforms (primarily Facebook and Instagram) to promote their products and services in the hair and beauty industries. In a gesture towards accountability, the Facebook Ad Library was set up to archive political ads posted to Meta platforms (Leerssen, et al 2023:1394). Commercial ads for hair products are not considered a “political issue” and thus they are not archived or accessible via the Meta API, but they are briefly available in the Library while active on University of Cape Town University of Cape Town Meta platforms. Our aim was to collect a sample of active campaigns over a period of time to document SA-targeted racialised messaging about hair on Meta platforms. We sought to explore (1) the range of adverts available via the Facebook Ad Library matching the keyword “hair” and which targeted Facebook users in SA (2) how the visual messaging of these adverts represented anticipated consumers and their hair, and (3) what such messaging might suggest about micro-targeting practices in the industry. Methodology A script was used to search the Facebook Ad Library using the query “hair” and restricting results to SA-targeted posts. The resulting purposive sample of ads (N=558) included 126 unique images and 85 unique video files posted to Meta platforms over a six month period by 99 different advertisers. Both quantitative and qualitative multimodal content analysis was conducted. This paper presents a visual content analysis of posts from the sample (n=183) which included images of people. When possible, signifiers of gender (feminine, masculine, multiple, not coded) were coded for any models and influencers featured in the images, along with a categorization of ethnicity (BIPOC, White, multiple, not coded) as suggested by the model or influencer’s appearance. Hairstyles were coded for length and texture or styling. These broad categories were used for intersectional analysis, and results were used to select case study campaigns for in-depth qualitative analysis. The long tail of a hairy industry The distribution of advertisements in the sample followed a long-tailed distribution (F (1, 15) = 21.37, p < .001), with most of the 99 advertisers in our sample posting only one or two advertisements. One local "hair growth treatment” brand (see images 1 and 2 in Fig 1 and image 2 in Fig2) was responsible for 160 of the sampled posts and 12 of the images. Figure 1: Distribution of SA-targeted advertisements (N=558) featuring the keyword “hair” on Meta (SA) ’23-‘24 Most ads were posted by smaller, relatively unknown local brands, including many individual hairstylists, hair salons and hair product manufacturers. Only one high profile consumer brand made an appearance in the sample. Visual analysis of gender and ethnicity Figure 2: Most prolific campaigns featuring WoC Content analysis showed that photographs of BIPoC women (see Figure 2) appeared in almost two thirds of posts (62%) making this the single largest category of posts. About half of such posts (32%) included signifiers of Black African identities, such as highly textured hair. Figure 3: Most prolific campaigns featuring White women In SA, White women are a tiny minority, but they were over-represented in our sample (20%, see Figure 3), much as they have been in SA media (Pilane, & Iqani, 2022). Far fewer ads featured men (9%), suggesting that the burdens and pleasures of hair care and maintenance are distinctly feminized while the hair industry is actively involved in production of gender. The only employment ads in the sample (for hairdressers) featured white men. Only 3 posts in total featured BIPoC men. Straight hair, #HairVirginity and purity discourse Images of long straight hairstyles appeared in the majority of posts (62%) and this style was idealized across ethnicities. By contrast, hairstyles associated with African ethnicities were underrepresented (11%). Straight, straightened and relaxed hair was associated with discourses of luxury, light, wealth, perfection, and purity. In these campaigns, textured hair was often described as dull, lifeless, frizzy or flawed. Some campaigns even implied relaxed hair was impure or tainted, such as the Perfect Angel post below (Fig 4), which associated wigs made from “virgin” hair (untreated human hair) with discourses of purity and love. Figure 4: Image and copy from ad for Perfect Angel wigs, retrieved from Facebook Ad LIbrary From messaging to targeting Meta does not provide data about the overall spend, reach or targeting of commercial ads. Hence we do not know how often these ads were displayed, nor to whom. Nonetheless the visual evidence in our sample suggests the presence of traditional binary demographic targeting by gender (Bivens & Haimson, 2016; Bivens, 2017). BIPoC women were targeted with messaging which, at best, promoted a new aesthetic or promised to save money or time. At worst, it was promoting harmful products and perpetuating racialized discourses. The ads in the sample were all associated with the generic term “hair”. It will be important to also explore more specific keywords (e.g. “braids” “Afro”) in future work, and to document possible outcomes of Meta’s decision to make terms like “Afro” unavailable for targeting to guard against them being used to discriminate against groups of users (Keegan, 2021). Given the distinctive messaging around White women and WoC and the examples we found of employment ad campaigns featuring only White men, we recommend further exploration of how certain Facebook advertisers may be using racial proxies to target by ethnicity (Keegan, 2021) by targeting Facebook-defined “interests” such as Afrikaans (language), or wigs, braids and hair straightening. Built on dataveillance, such micro- targeting may be fueling feedback loops which intensify harm or deepen existing gendered inequalities. Conclusion Despite the limitations and restrictions of the Facebook Ad Library, the sample yielded many insights into how this industry may be adopting targeted ads. Our analysis of the messaging about hair in the ads showed racial ideologies and binary gendering at work in campaigns where the visual imagery addresses women in SA. This initial analysis also suggests ways to use active campaigns on the Ad Library to reduce the obscurity surrounding algorithmic targeting on social platforms. A range of other approaches (see e.g. Angus, et al, 2024) will help engage South Africans about the ads they see online, begin documenting potential harms, and help demand accountability of Meta platforms in the Global South. List of References Angus, D., Hayden, L., Obeid, A. K., Tan, X. Y., Carah, N., Burgess, J., Parker, C., Andrejevic, M., Fordyce, R., Cellard, L., & Bagnara, J. (2024). Computational Methods for Improving the Observability of Platform-Based Advertising. Journal of Advertising, 1– 20. https://doi-org.ezproxy.uct.ac.za/10.1080/00913367.2024.2394156 Bivens, R. (2017). The gender binary will not be deprogrammed: Ten years of coding gender on Facebook. New Media & Society, 19(6), 880–898. https://doi.org/10.1177/1461444815621527 Bivens, R., & Haimson, O. L. (2016). Baking Gender Into Social Media Design: How Platforms Shape Categories for Users and Advertisers. Social Media + Society, 2(4), 205630511667248. https://doi.org/10.1177/2056305116672486 Bucher, T. (2017). The algorithmic imaginary: Exploring the ordinary affects of Facebook algorithms. Information, Communication & Society, 20(1), 30–44. https://doi.org/10.1080/1369118X.2016.1154086 Elias, A. S., Gill, R., & Scharf, C. (2017). Aesthetic labour: Rethinking beauty politics in neoliberalism. Palgrave Macmillan UK. Keegan, J. (2021, July 9). Facebook Got Rid of Racial Ad Categories. Or Did It? – The Markup. https://themarkup.org/citizen-browser/2021/07/09/facebook-got-rid-of-racial-ad- categories-or-did-it Madlela, K. (2018). Black hair politics: The representation of African women on True Love magazine front covers and hair advertisements [Thesis, University of Pretoria]. https://repository.up.ac.za/handle/2263/65573 Sobande, F. (2020). The Digital Lives of Black Women in Britain. Palgrave Macmillan. Spector, M., Naidu, R., & Cooke, K. (2023, November 4). Thousands of Black women claim hair relaxers gave them cancer. Reuters. https://www.reuters.com/investigates/special-report/loreal-hair-relaxers-litigation/ Tate, S. (2007). Black beauty: Shade, hair and anti-racist aesthetics. Ethnic and Racial Studies, 30(2), 300–319. https://doi.org/10.1080/01419870601143992