Ad frequency is impressions divided by reach: how many times, on average, each person in your audience has seen the same ad. If a hypothetical campaign serves 30,000 impressions to 10,000 people, its frequency is 3. Some repetition is useful — most people do not act on first exposure — but past a point repetition curdles into fatigue: the audience stops noticing the ad, then starts resenting it.
Fatigue shows up in the metrics as softening CPM efficiency and climbing costs, but it shows up first, and most vividly, in the comments. A fresh ad's thread is questions and tags; a fatigued ad's thread turns sarcastic — "stop showing me this," "I see this every day," jokes at the creative's expense. Because comments persist on the ad post, this snark becomes part of the ad experience for everyone who sees it next, compounding the fatigue it reflects. High frequency also gives annoyed users more chances to hide or report the ad, feeding negative feedback signals that make delivery more expensive.
That makes the comment section one of the earliest fatigue detectors you have. Dashboards tell you frequency crossed a threshold; the thread tells you how the audience feels about it, often days earlier. A moderation workflow that classifies comment tone — the job of sentiment analysis — can flag when an ad's conversation is souring, which is a strong hint it is time to rotate the ad creative, refresh the audience, or cap frequency.
For an ad-comment tool like ROAS Shield, frequency is context: the same complaint means something different on an ad the commenter has seen once versus one they have seen a dozen times. Reading comments ad-by-ad, with campaign context attached, is what turns a pile of replies into an early-warning system for tired creative.