The product was fine; the app it shipped with was the problem

Seventeen percent of units came back, and two thirds of those returns happened within three days of delivery — far too fast to be a durability problem. The hardware tested clean; the returns were happening before anyone had really used it.

The cause was the setup. The gimbal does nothing until it is balanced, paired and calibrated through a companion app, and that sequence took an inexperienced buyer somewhere between twenty and forty frustrating minutes, with the app's own instructions written for someone who already knew. A customer who cannot make a $119 device switch on in the first evening does not persevere — they return it and leave a review saying it does not work.$

We treated the first fifteen minutes after unboxing as part of the listing.

• A setup sequence rebuilt as three steps, each one verifiable by the customer, replacing an eleven-step flow with no feedback.
• The balancing step demonstrated on the detail page before purchase, so the buyer knows it exists and is not ambushed by it.
• A short first-run video reachable from the packaging, showing the exact sequence with the exact screens, in the three market languages.
• Review text routed back into the flow monthly, so each new confusion point was fixed rather than argued with in a reply.

Returns fell from 17% to 5%, and the share of returns happening in the first 72 hours dropped from 68% to 19% — the specific failure had been removed rather than displaced.

The rating rose four tenths to 4.2, because the reviews that had been scoring the setup experience as a hardware fault largely stopped being written. Net revenue grew 93% against 58% growth in gross orders: roughly a quarter of the gain is not new demand at all, it is orders that used to come back.