What listeners actually hear
Faster speech persuades - until it doesn't
Miller, N., Maruyama, G., Beaber, R. J., & Valone, K. (1976). Speed of Speech and Persuasion. JPSP. See also Smith & Shaffer (1991) and curvilinear-rate meta-analyses (Kim et al.).
Miller and colleagues found that speaking faster made a message more persuasive and the speaker more credible - listeners read speed as competence. But "more is better" is the wrong lesson. Later work showed the relationship is an inverted-U: push past roughly 190 words per minute and comprehension falls off a cliff. The target is not "fast." It is a band.
Why a band, not a number#
Most delivery targets are curvilinear - there is a cost to too little and too much. Coaching that rewards "ever faster" pushes speakers straight past the peak into incomprehension. The right design is a dead-band: a tolerance zone where the dimension is "good enough" and the coach stays quiet, flagging only genuine drift outside it (bandwidth feedback, Sherwood 1988).
What it means for Speech Away#
We compute pace from Deepgram's word-level timestamps and score it against the 140-160 band, flagging both sustained too-fast and too-slow - never rewarding raw speed. And because the fix for runaway pace is the same as the fix for fillers - a one-second silent pause - the coach often hits two dimensions with one cue.