How a DTC skincare brand recovered $91k/month in ad spend with automated comment moderation

GlowLane runs a portfolio of Meta Ad campaigns promoting its DTC skincare line. Creative, audience, and bid strategy were all locked — the team had tested extensively — and for 18 months the campaigns performed at a consistent 3.4x ROAS.

Starting in Q4 2025, ROAS began drifting down. Nothing in the media plan had changed. The creative rotation was fresh. The audience definitions were stable. But the numbers kept sliding — by late January, blended ROAS had dropped to 2.6x and the CFO was asking questions.

The performance team audited the funnel: landing pages, checkout flow, product copy, ad bidding strategies. Nothing explained it. The only variable they had not instrumented was the comment section under each ad.

FeedGuardians ran a 7-day read-only audit across GlowLane's active Meta ad campaigns. The results were unambiguous. 19% of all comments under active campaigns were harmful — scam replies, competitor bait ("Brand X is cheaper and better"), and complaint dogpiles under ads that had gone slightly viral.

The most damaging pattern: the team had a prior quality control incident six weeks earlier, and customers who had been refunded were still leaving angry comments under any new ad creative. Every time GlowLane launched a new campaign, the same 30 detractors showed up within hours, leaving visible warnings to potential buyers.

The CTR penalty was measurable. Impressions served after the first negative comment was visible had CTR 26% lower than the same creative served to audiences that saw a clean comment section.

GlowLane deployed FeedGuardians in Standard mode — automatic hiding of spam, scams, profanity, and competitor mentions. They added a custom rule set for their specific competitor brands and for the specific keywords tied to the old quality control incident.

For legitimate customer questions, the team turned on AI auto-reply trained on GlowLane's FAQ. Most questions under ads were about shipping times, sensitive-skin suitability, and return policy — all answers the AI could handle in under 2 seconds.

Escalations for edge cases went to a Slack channel shared with the brand and community managers. About 3% of flagged comments needed human review; the AI handled the other 97% automatically.

Week 1: harmful comment visibility dropped from 19% to under 2%. The team noticed CTR on new creative rounds beating the old cohort baseline for the first time in months.

Week 3: blended ROAS recovered from 2.6x to 3.1x. The performance team correlated the lift directly with the comment-health improvements in their FeedGuardians dashboard.

Week 6: ROAS peaked at 3.95x — 51% above the pre-moderation baseline of 2.6x and 16% above the brand's all-time best before the decay started. Media spend was held constant throughout so the lift was purely attributable to the comment-side change.

With $400k/month in ad spend and a 51% ROAS lift, GlowLane recovered approximately $91k per month in ad spend efficiency — a direct line to bottom-line profit since ad spend was held constant.

Over the first six months of the engagement, cumulative recovered ad spend surpassed $540k. The FeedGuardians subscription cost $89/month across all connected accounts. ROI was approximately 1,000x.