What happened when we put real Aera customers on camera instead of actors.

[ 01 ] · What we tried

Six customers. Six hooks. Same brief.

We pulled six Aera customers from a recent post-purchase survey. Skewed for variety: different rooms, different routines, different reasons they bought. Sent them a short brief, a shot list, and a few prompts.

Each one filmed a 60-second response on their phone. We did one round of light direction over text. No re-shoots. No script. We took what they sent and cut it against the same hook patterns we were already running with our scripted variants.

To make it a fair test, we held the rest of the ad constant. Same product b-roll, same end card, same offer. Only the talking-head opener changed.

[ 02 ] · The result

The good ones beat the scripted variants. The bad ones bombed harder.

Higher variance than scripted variants in both directions. The two winners outperformed our best scripted hook for six straight weeks. The three losers had a lower hook rate than even our weakest scripted variant. One landed in the middle.

A real customer with the wrong delivery is worse than an actor reading a tight script. A real customer with the right delivery is better than anything you can write.

[ 03 ] · What worked

Specificity. Always specificity.

The winning hooks all named a specific moment, room, or routine. "I started using it in the bathroom because my husband hated the smell of the candles I was burning." "I run this in the nursery during nap time because the diffusers are too loud." Concrete, weird, clearly real.

The hooks that died all sounded like a marketing line. "I love how it makes my home feel." Even from a real customer, that phrasing reads as scripted. The audience can tell.

[ 04 ] · What didn't

Light direction wasn't enough for half of them.

We assumed real customers would naturally talk like real customers. Half did. Half slipped into review-speak the second a camera was on, repeating the brand's marketing language back at us. We needed a tighter intake: a 15-minute call before they filmed, asking specific stories about specific days, not asking why they liked the product.

We also under-invested in framing and audio. Two hooks tested well in dailies but landed in the bottom quartile in-feed because the audio was rough or the framing was distracting. Phone-shot is fine. Phone-shot with a $30 lav mic is meaningfully better.

[ 05 ] · Next time

Three changes for the next round.

• Story-mining call before the shoot. A short call to surface the weird, specific moments. The hook is in the conversation, not in the brief.
• Ship a small kit. Lav mic, a clip-on light, a one-page framing guide. Lifts the floor without losing the authenticity.
• Go bigger on volume. Six was too small. Twelve to fifteen lets us cull harder and find more winners.

[ 06 ] · When to use this

If you have an in-house buyer running weekly tests, this is a one-week experiment worth the slot.

It's cheap to set up, it gives you a clear read in two weeks, and the winners can carry rotation for months. The losers also teach you what your scripted hooks are doing that customers can't fake: tight pacing, rhythm, an actual reason to keep watching.

If you don't have a tight script you trust as the control, run that first. You can't read this signal without one.