55 days. ROAS 34.
The problem was never the ad or the creative. The problem was that the owner was content.
When the account was handed over he was running ads from the app and nothing else. No data, no record of who bought, no system. He wouldn't spend on UGC ("I'll get an employee to shoot it on a phone"). He didn't want a website even at our expense.
The first move was creative people actually believe. The product was tested by hand and studied against competitors, and the videos were made deliberately raw: employees and ordinary people filming as though posting to their own accounts. "I bought this and I'm going to try it with you." People have memorised the look of a traditional ad and the look of polished UGC, and they have started trusting background noise and an unshielded mic instead.
The start was conventional. TikTok was tested with no real effect and we returned to Facebook, because the owner refused to wait it out. Many videos were tested, a serious amount was spent, and the return was poor. Until one video landed. Messages arrived at a volume that made the moderation team ask for help, so a bot was connected to move comments into messages with a moderator finishing the conversation. That was the first thread.
Next was building on whoever buys. Back to the owner: a website? No. Instagram? He disliked it, and after some persuasion agreed, having never had an account. Budget rose gradually. Bundles and offers. Follow-up with customers for reviews and unboxing videos in exchange for a gift, a discount, or cashback. The goal was raising sales without damaging ROAS, while improving delivery rate and units per order.
Then came the moment that opened the door. We wanted to scale and he refused, because the numbers were fine and he was satisfied. Until a competitor appeared in another governorate selling cheaper, and customers could cancel and buy there. As the numbers slipped, he decided to increase spend and gave full authority.
We scaled on winners: every winning creative pulled out and scaled against the data. Instagram started working. Offers and bundles lifted AOV, so budget rose gradually every day and a half to two days with close monitoring.
After exactly 55 days: more than 60,000 messages, 25,502 orders with 16,576 delivered (above 65%), ROAS 34.
Success isn't opening an ad account. It's knowing when to expand, and being able to predict what happens next.
The lesson
Data before ads
