Social Rocket
e.l.f. Cosmetics combines a staged witch comedy with Reels, TikTok and YouTube in "The e.l.f.ing Magical Sisters". On 30 September 2026 the format went online with Priah Ferguson, Keyla Monterroso Mejia and Bebe Wood. The brand now reports first engagement signals. For social media marketers, however, what matters more is which metric actually says something about product sales – and which does not.
A campaign starting with entertainment
The film promotes the Soft Glam Cream Blush & Bronzer Sticks, but initially has the three actresses appear as characters in their own comedy format. The campaign comes from the brand entertainment unit e.l.f. Made. Those involved include writer Anna Drezen and director Hannah Levy, among others. The company announcement of 30 September confirms the cast, products and distribution via Instagram, TikTok and YouTube.
The dramatic trick: the clip works as a small entertainment story, and only afterwards does product performance – easy application and blendability – become the motif. The borrowings from the series and film aesthetics of the 1990s and 2000s are a creative device. The bigger marketing goal is the reusability of a world of characters, language and product moments across several social platforms.
What the numbers prove – and what they don't
According to e.l.f., the main video was the Instagram post with the most interactions in September and also reached the brand's highest engagement rate for the month there. Marketing Dive reported on this on 23 October. Neither an absolute interaction figure nor a comparison of the user groups reached nor an independent brand lift measurement was cited. The statement is therefore an internal month-on-month comparison, not proof of higher advertising impact than other campaigns.
Even more important is the timing separation in sales: e.l.f. explains that 50 percent of the buyers of the advertised blush and bronzer sticks had not previously been buyers of the brand. This company figure explicitly refers to May to September 2026. Because the campaign film only appeared on 30 September, no new-customer effect caused by the film can be derived from it.
What a useful effectiveness test would look like
For comparable creative variants, brands could separately measure how many people watch the clip up to the product moment, how the search for the product changes, and whether additional buyers are gained. For a reliable claim of impact, test and control groups, clean time windows and a separation of paid and organic reach would be required. Comment requests for more episodes show interest in the characters – not automatically purchase intent.
e.l.f. has been experimenting with its own cultural formats for some time: the album project "Mirror Mix" combines music, Roblox and live appearances. In social gaming the brand also relies on ZEPETO beauty looks. Another route to a recurring brand format is described in our analysis of Banana Republic's TikTok series "The Agency".
Classification: The new campaign offers an example of how a brand earns attention for a product with its own entertainment. Whether the production actually brings additional sales is open based on the published data. For German marketing teams, the learning lies less in the witch aesthetic than in the clean separation of creative success and business impact.
Sources: e.l.f. Beauty, official press release of 30.09.2026 · Marketing Dive, interview with e.l.f., 23.10.2026.
Stock image: Lana Mattice / Unsplash; not a scene from the e.l.f. campaign. The official company press image could not be technically adopted.