Weekly spend from $3.3K to $8.8K. ROAS held at 2.57.

More spend broke the account.

In June and July the brand pushed Meta to $8K to $9K a week. ROAS fell to between 1.1 and 1.9. By late August they had cut back to about $3K a week.

The ads were lifestyle shots and licensed characters with almost no words. Nice to look at. No reason to buy.

Buyers don't buy a smell. They buy a room.

I went through the brand's 8,262 reviews. Of the 2,000 tagged by theme, half (49%) talked about the experience: what the candle does to a room and a night.

“It truly made my home feel like an elegant, mysterious, inviting gothic cathedral.”Review of the lead candle

“Smells like dark velvet and fall colors in a witchy room.”Review of the second candle

Two kinds of buyer came out of it. One buys a mood for the room. The other buys a piece of who they are. Both show up under every candle, so I wrote for the buyer, not the product.

Six concepts for one candle, then 20 test batches.

I briefed six static concepts for the lead candle, each testing one idea: a review as the whole ad, us against a drugstore candle, a “75% complete” Halloween checklist, and more. The brand ran them in 20 batches across the month.

One headline came straight from a review. A buyer wrote that guests kept asking “what is that amazing smell” and that “it literally drew people in”. That line became an ad.

Three of the ads

2.7 times the weekly spend, same ROAS.

Over the 30 days the account spent $27,566 on Meta and brought $67,164 in sales. The week-by-week view matters more than the total:

From the brand's own daily tracking sheet, Facebook columns.

Across the month, the lead candle's ads returned 3.58 ROAS on $8,614. The second candle's returned 2.43. That gap told the brand where to put the money.

What I don't count: one review-quote ad hit 9.03 ROAS, on 7 purchases and $68. That is a signal, not a winner. The ad that carried the money was the “This is atmosphere” concept: $6,760 spent at 3.14.