We read 1,495 Instagram comments across three DTC niches. 72% was noise. The rest was money.

This week we pointed our pipeline at 14 brand accounts across three DTC niches — skincare, daily-wellness supplements, and running gear — and read every comment on their recent top posts: 1,495 comments in total. Every number below comes from that dataset. We publish aggregates and patterns only; we never republish individual comments or identities.

Finding 1: 72% of comments are noise — which is why nobody reads them

Of 1,495 comments, 411 carried signal after filtering emoji chains, duplicate hype and spam. That 72% noise floor is the honest reason brands don't read comment sections: the ore is real, but nobody has time to sift it by hand. It is also why the signal goes unread by the brands it is addressed to.

Finding 2: purchase intent sits in public, unanswered

73 comments were explicit buying signals — “need this”, link requests, drop-timing questions, purchase confirmations. In skincare they were 23% of all signal: a launch-week comment section functioned as a public pre-order queue. Almost none received a reply from the brand.

• Skincare: 54 purchase-intent comments — the launch-and-restock niche.
• Running gear: 19 — clustered on a single collab drop.
• Wellness: near zero — a giveaway week produced volume, not buyers. Format shapes the signal you get.

Finding 3: negative comments are rare — and worth the most per word

Under 4% of analyzed comments were negative. But that thin slice held the findings a founder would pay for: a subscriber describing how cancelling felt impossible (a churn machine running in public), and a reformulation thread where relief and resentment arrived together — audiences tolerate flawed products, but they remember the wait. If you read only one kind of comment, read the angry ones.

Finding 4: feature requests are free product research

28 comments were direct requests: colorways, a dropper redesign extended to a second product, the ability to pause a subscription instead of cancelling it. None required a survey, a panel, or an agency — they were sitting in competitors' comment sections, timestamped and like-ranked.

What this means if you run a DTC brand

• Your competitors' launch posts are demand data for your inventory and content decisions.
• Sub-4% negativity means every complaint deserves triage — it is the densest signal you have.
• If your engagement strategy is giveaways, know that you are buying volume, not buyer intent.
• Answer the questions people ask in comments in your captions — almost nobody does.

Method note: one week's top posts per account, 14 accounts, three niches, comments classified by the same pipeline that builds our weekly briefs. Small sample, honestly labeled — we'll publish bigger multi-week cuts as the data accumulates.

Want this run on your niche? A weekly Lurkfish brief reads thousands of comments across the accounts you choose — every claim linked to the verbatim comments behind it. Or start smaller: the free post autopsy analyzes any single Instagram or TikTok post.