Consumer Skincare

A premium skincare brand with strong retention and a familiar consumer problem: acquisition costs were rising, influencer content was saturated, and buyers were doing more research before purchasing. In skincare, Reddit is the decision engine, and the brand was not being attacked, it was being omitted. Omission is harder to fix than attack because you do not get a chance to correct the story.

At engagement start, brand presence sat at 5 of 71 tracked Reddit URLs (7%). In the first 14-day reporting window, 1 AI chat answer named the brand, in a long list and in fifth position.

We treated this as a credibility problem, not a posting problem. In skincare, readers want evidence, mechanism, and practical usage advice. Lifestyle marketing does not get cited; specifics do.

Posts and in-thread responses were rooted in formulation logic, realistic expectations, and clear guidance on who the product is not for. The 'who this isn't for' framing is what survives a skeptical comment section and becomes quotable later. We engaged with every meaningful pushback in the comments, including the bad-experience ones, without trying to erase criticism.

Timing mattered. Skincare buying conversations surge around seasonal changes, sale cycles, and product launches. Posts published during those windows are more likely to be saved, shared, and resurfaced later when search and AI synthesis pull from the period.

Paid amplification ran only on angles that had already proven themselves organically. The ad creative pointed to educational pages designed to be useful in search results indefinitely, not promotional landing pages designed to convert once.

Engagement ran for 120 days, structured as 8 rolling 14-day reporting cycles. Each cycle produced the same standardized report we use across our engagements.

Brand presence counts tracked Reddit URLs (URLs AI chats pulled as sources in the period) where the brand was either named alongside the URL by the AI, or where we had placed a comment at any point in history.

AI answers naming the brand is taken from a third-party AI search monitoring layer running a fixed prompt set against ChatGPT, Gemini, and Google AI Overviews daily. The Reddit-citation subset is computed by inspecting each named-the-brand chat and checking whether its sources include a reddit.com URL. The 'cited a thread we comment in' subset is the same calculation joined against our internal posted-comment database.

Category ranking is from the brand's main retail partner's dashboard, reported per month. Branded search lift is from the client's search console, measured after the high-signal threads went live.

What this generalizes to: consumer categories where Reddit threads regularly rank on page one for high-intent purchase questions. In those categories, being omitted from a thread is functionally being omitted from page one and from the AI answer that summarizes the thread. The standardized cycle report makes that visible cycle by cycle, which is what allowed the client to stop treating Reddit as a brand-safety risk and start treating it as a search surface.

What this does not prove: that the same arc works for a category where the buyer does not pressure-test the product before purchasing. Impulse-purchase categories will not see the same AI answer movement, because LLMs cite where the public discussion lives. If the public discussion does not exist, the program builds it from scratch and the timeline is longer.