Formats1 min read

    Does a voice note sell for more than a written answer?

    Across 548 paid answers, the median voice note sold for $27 and the median written answer for $18. But the average written answer was higher than the average voice note, which tells you something more useful than either number.

    $27

    median voice note

    $18

    median written answer

    548

    paid answers read

    Who is writing this: Lucy Chen, who runs Nudge. I read these figures out of our own payments rather than collecting them from anywhere else. Captured payments joined to their questions, grouped by answer format, 2026-10-18. Medians and means are both reported because they disagree.

    Who this is for: coaches setting a price list, and anyone about to make text the cheap tier.

    The medians say voice

    Voice: 271 paid answers, median $27. Text: 263 paid answers, median $18. Video: 13 paid answers, median $57. A typical voice note sells for half again what a typical written answer sells for, and video for three times it.

    The averages say text

    Average paid: text $32.04, voice $29.51. Text is lower in the middle and higher on average, which only happens when a handful of written answers sold for a great deal more than the rest. The long tail of text is fatter than the long tail of voice.

    What that actually means for a price list

    Voice is the better default price. Text is where the unusually large tips happen, probably because a written answer is what someone asks for when the question is detailed enough to be worth a lot. Pricing text as the cheap tier caps that.

    The rating surprise

    The assumption is that a voice note feels more personal and therefore rates better. It does not. Written answers average 4.84 from 92 ratings, voice 4.75 from 114. The gap is small but it runs the opposite way to the folklore.

    Video is the format almost nobody offers

    Thirteen paid video answers, median $57, and every one of the three that was rated got five stars. That is far too small a sample to prove anything, which is exactly why it is worth testing rather than citing.

    Where this is weak

    One legacy row is stored as "audio" rather than "voice" and is excluded. The video numbers rest on 13 payments and 3 ratings, so treat them as a prompt to try it, not as evidence.

    More from the data

    Put a price on the answers you already give

    Every figure in these posts came from coaches doing exactly that.