How we took 102 hours a month of manual campaign work off a premium organic skincare brand selling into 25+ markets

Challenge

The brand came to us scaling fast, with five problems in its paid acquisition:

• The manual workload had become unsustainable: optimising campaigns by hand was eating the in-house team’s week.
• Difficulty managing larger budgets: as spend increased, so did the complexity of maintaining ROAS and controlling costs.
• Limited ability to react quickly: with nobody on the accounts daily, performance changes were caught late.
• No scalable system: running 20+ ad accounts with no single operating standard made optimisation inconsistent from market to market.
• Risk of wasted spend: delayed decisions and manual errors increased the likelihood of inefficient budget use.

Solution

The brand hired us to run paid acquisition on Meta and TikTok. This is how we work the accounts:

‍

ROAS-first buying

We agree a ROAS target per account and buy to it, holding performance inside that range market by market.

‍

Optimisation at every level

We work the campaign, ad set and ad level separately, moving budget and bids on what each level is actually producing.

‍

Day-part optimisation

The buying approach is the same everywhere, but we run it on a schedule built per market — country-specific buying behaviour and peak hours.

‍

Bid management

We adjust bids daily, both to protect margin and to push spend into what is working.

‍

Delayed-attribution recovery

We monitor ad performance daily and re-check paused ads against a five-day window, restarting the ones whose ROAS and conversions landed late.

‍

‍

Results

• Took up to 102 hours a month of manual campaign work off the in-house team
• Brought 20+ ad accounts onto one operating standard, which held performance consistent across markets.
• Cut wasted ad spend by pausing ads that fall under their ROAS threshold
• Scaled winning campaigns the same week, without waiting on a monthly budget review
• Made faster, evidence-led decisions — we monitor ad performance daily and review it against custom performance windows