Last quarter we ran an audit across 312 small-business Meta Ads accounts that ran >€500/month in spend. We measured one specific number: what percentage of comments under paid ads contained a price, availability or contact question — and what percentage of those got a reply within 4 hours.
The headline: 30.4% of comments were buying-intent. Of those, only 27% got a same-day reply. The other 73% — purchase intent that the brand paid Meta to surface — silently dropped out of the funnel.
This article walks through the playbook to fix that. Manual baseline first (so you understand the mechanics) and then automation (so you can scale).
Step 1: Measure your current leak
Open Facebook Ads Manager. Filter to last 30 days. Pull the top 10 ads by spend. For each ad, scroll through every comment and tag them: lead, question, spam, hate, praise, other. Count the leads. Count which got a reply within 4 hours.
The ratio (replied-leads / total-leads) is your leakage baseline. We've seen anywhere from 5% (terrible) to 60% (good for a small shop with one engaged owner).
Step 2: Tighten the manual process
Three rules:
- Set a daily 09:00 + 18:00 slot for comment triage. 20 minutes each.
- Reply to every lead with the same template: "Hi [name]! Thanks for asking. Sent you a DM with full details." Then DM the actual answer.
- Hide every hate / spam in the same slot. Block repeat offenders.
This alone should bump your leakage from 27% → 50-60% within two weeks. The cap is human bandwidth — 40 minutes/day moderating comments is hard to scale beyond €3 000/month in ad spend.
Step 3: Automate the obvious
For any ad spend above €1 000/month, manual moderation stops being economical. The math:
- 40 min/day × 30 days = 20 hours/month moderating comments
- At €25/hour your manual cost is €500/month
- Lead recovery from manual moderation: ~50% × N leads
- An AI moderation bot at €25-50/month does 95% recovery
So the break-even is around €100/month in saved labour OR 5 additional recovered leads. Whichever you hit first.
Step 4: Pick the right tool
The decision tree:
- You want comment-specific moderation with auto-DM: Mochu (full disclosure: that's us).
- You want general DM funnels with visual flow builder: ManyChat.
- You want enterprise multi-channel scheduling + analytics: Sprout Social.
- You want pure scheduling, no moderation: Buffer, Later, or the native Meta Business Suite.
Step 5: Onboarding done right
Whatever tool you pick, the onboarding pattern is the same:
- Connect via OAuth. Never share passwords with a third-party SaaS.
- Start in semi-auto mode for a week. Approve every bot reply manually. You'll spot tone problems early.
- Lock in industry settings. A beauty salon needs a different default tone than a B2B SaaS. Most tools have presets.
- Audit weekly for the first month. Skim 20 random bot replies. Adjust custom instructions if needed.
- Move to full-auto in month 2. You'll know enough to trust the bot for the routine cases; keep semi-auto only for ambiguous categories like 'other'.
Step 6: Measure the recovery, not the bot
Don't measure "how many comments did the bot reply to". That's a vanity metric. Measure:
- Reply latency P50 — median time from comment posted to bot reply (should be <5 seconds).
- Lead recovery rate — % of lead-classified comments that turned into a real conversation (DM exchange of 2+ messages).
- Cost per recovered lead — (bot subscription) / (recovered leads). Should be under €10 for a healthy SMB account.
- Customer-experience score — survey 5 customers who got a bot reply. "Was the response helpful?" — target 80%+ helpful.
Common mistakes
- Setting and forgetting. Bots need a quarterly check-in. Industry context drifts, your offer changes, the bot needs a tune-up.
- Going full-auto from day 1. Always start semi-auto. Trust must be earned.
- Picking based on price alone. A €15/month tool that requires 5 hours of setup is worse than a €50/month tool that's running in 5 minutes.
- Ignoring the hate / spam half. Your CTR depends on the comments section looking clean to other prospects.
What to do this week
If you're spending €500+/month on Meta Ads and don't have any automation: try Mochu free for 14 days. Measure your baseline before installing. Measure again 30 days later. The number is usually surprising in the right direction.
If you're spending under €500/month: skip the bot, but set the 09:00 + 18:00 manual moderation rule from Step 2. That alone moves the needle.