Orders up 256% by finding the ad set that actually worked

Find the ad set that works, then feed it

BabyFits sells baby clothing through Shopify. The account was already spending on Meta, but spend was spread thinly across a large number of ad sets with no clear read on which audience or creative was actually producing orders. The work was to find the combinations that convert and move budget behind them.

Client

BabyFits

Year

2026

Industry

E-commerce · Baby & Kids

Expertise

Paid Social, Creative Testing, Shopify, Conversion Tracking

Platform

Meta · Shopify

Eighty-nine ad sets were live. At that spread, no single set accumulated enough data to prove itself, cost per purchase varied by a factor of five between them, and creative fatigue was invisible because nothing ran long enough to fatigue. Broad-audience testing on Meta only works if you let a set run, which means being willing to turn most of them off.

Structured testing against broad audiences rather than narrow interest stacks, with product-led creative and enough budget per set to reach statistical daylight. Winners were identified on cost per purchase rather than on CTR or reach, since a cheap click that does not convert is just a cheap click. Budget then consolidated behind the sets that held their cost as spend increased, with the Shopify side tracking order volume and average order value so the ad platform was never the only source of truth.

Across Jul 1 to Aug 19, 2026, the store recorded PKR 1.05M in total sales from 770 orders, up 163% and 256% respectively on the preceding period. On the Meta side, the strongest broad ad set delivered 113 website purchases at PKR 635.61 each across the 30 days to Aug 18, against 1.53M impressions and 708,501 people reached.