Get your free personalized podcast brief
We scan new podcasts and send you the top 5 insights daily.
Viral marketing firms now focus on manipulating recommendation algorithms rather than appealing to human taste. They spam platforms with content to trick the algorithm into believing a song or product is already popular, making the algorithm the primary customer, not the end user.
Related Insights
Unlike older algorithms that recommend content based on long-term follow history, TikTok's model prioritizes recent engagement. This 'TikTokification' across platforms means algorithms can now find an audience for off-niche content if it aligns with a viewer's immediate, short-term interests.
Modern algorithms can surface any single piece of content to a massive audience of non-followers, regardless of past performance. This means marketers are always just one breakout post away from significant reach, making consistent experimentation more important than ever.
Recommendation algorithms don't just predict what users like; they actively nudge users toward more extreme preferences. This makes behavior easier to predict and monetize, effectively creating an automated radicalization pipeline for the algorithm's own efficiency.
The traditional goal of winning hearts and minds is now a two-step process. Marketers must first win over the "machines"—search algorithms and LLMs—that control 85% of content discovery, treating them as an influential, gatekeeping audience.
A new marketing tactic involves creating high-quality, AI-generated content on platforms like Reddit to promote a product. The goal is to have this seemingly authentic user content indexed and then surfaced by LLMs like ChatGPT in their summaries, creating an insidious and hard-to-detect marketing channel.
Stop thinking of 'feeding the algorithm' as a cynical game. The algorithm's core function is to surface what people find relevant. Therefore, creating content for the algorithm is fundamentally the same as creating content that captures genuine consumer attention and serves their interests.
The 'industry plant' concept is now a calculated strategy. Firms use vast networks of social media accounts to fabricate interactions, share clips, and stoke discourse, effectively 'simulating a trend' to push a band like Geese into the recommendation algorithm and create artificial buzz.
Algorithms increasingly serve content to non-followers based on their interests, not just social connections. To succeed, marketers must shift from engaging existing followers to creating "recommendable" content that appeals to a broader, topic-focused audience.
Platforms like TikTok and Instagram no longer prioritize your social network. They are "interest algorithms" that surface the best content regardless of follower count. This increases content performance variance, making daily quality paramount over historical reputation.
Despite the dominance of platforms like Spotify, there's a growing fatigue with algorithmic recommendations. Consumers feel this approach can be impersonal and lead to a "lowest common denominator" experience, creating a market opportunity for brands that offer authentic, human-led taste-making and curation.
