An apparel brand's Advantage+ shopping campaign got rules, and stopped eating the good campaigns

Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.

Where they started

Seasonal drops anchor the apparel brand's calendar — four collections a year, each with a launch window the merchandising team plans months ahead. The business sells direct through its own store, paid social is the primary acquisition channel, and a small retargeting program was the team's pride. The automated shopping campaign had been adopted early and enthusiastically; by the time we arrived it consumed the majority of budget by default while the drops' prospecting plans quietly thinned.

What it was costing

Advantage+ shopping took 60% of budget by default and cannibalized the brand's retargeting and prospecting campaigns — its reported ROAS looked fine because it was harvesting demand the other campaigns created.

What they could see

• Advantage+ consumed most of the budget by default and its reported ROAS sat comfortably above every manual campaign's.
• Prospecting for each seasonal drop launched underfunded, then begged for budget through the quarter as the automated campaign held its share.
• Retargeting's reported numbers kept climbing even as its reach definition loosened — more of it was catching people already mid-checkout.
• When the team capped Advantage+ temporarily, prospecting recovered visibly within days, then lost ground again as budgets were 'rebalanced' back.

The constraints we worked inside

• The brand wanted automation kept — manual-only was a step backward they refused.
• Seasonal drops needed guaranteed prospecting budget — Advantage+ couldn't be allowed to absorb it.
• Incrementality was the honest question — platform ROAS for retargeting is known fiction.

What had been tried before

Budget was capped on the automated campaign whenever someone noticed prospecting throttling.

Manual caps fought the automation daily, fluctuated with whoever noticed, and never guaranteed the seasonal prospecting budget the drops actually needed.

Retargeting budgets were increased repeatedly to defend the segment the automated campaign was absorbing.

More retargeting spend harvested the same near-converting users at rising cost while the prospecting line — the actual growth engine — stayed starved.

A fully manual campaign structure had been trialed the year before we arrived.

Manual-only lost the automation's delivery efficiency, demanded daily hands-on management nobody had time for, and was abandoned within a quarter.

What we proposed

We proposed keeping Advantage+ but giving it a fence and a quota. Prospecting, retargeting, and the automated campaign each received defined budget shares with exclusions set so each fished in its own pond — the seasonal drop's prospecting budget became a guarantee rather than a residue. The biggest claim on the books, retargeting's platform ROAS, went under a holdout test to measure what it actually adds. Quarterly re-balancing runs on blended MER and the holdout findings, not platform ROAS comparisons.

Just as important is what we ruled out, and why:

• Switching off Advantage+ entirely — The brand had already tried manual-only and refused the regression; the automation's delivery efficiency was real — the failure was unbounded budget claims, not the machine itself.
• Data-driven attribution settings alone — Attribution models reshuffle credit inside the platform; they cannot answer whether retargeting adds revenue, which is a question about behavior, not bookkeeping.
• Consolidating everything under one Advantage+ umbrella — One umbrella campaign repeats the original problem at larger scale — budget shares would still be decided by the algorithm, just without visibility.

How the work ran

Delivered by the growth pod — paid social specialist over 6 weeks, with working increments reviewed with the client every week.

The stack, and the reasoning

Advantage+ shopping

Retained under a budget quota — the brand refused a manual regression and the automation performs once its cannibalization is fenced by exclusions.

Campaign exclusions

Audience and brand exclusions kept the three lines from bidding into each other's audiences, which is what made the budget shares meaningful at all.

Holdout incrementality test

Platform ROAS comparisons were the fiction under review; a holdout measures what a line adds rather than what it claims.

Blended MER reporting

Quarterly allocation runs on blended efficiency, which the finance contact can verify, instead of per-campaign ROAS the platform grades itself on.

What went wrong

How we worked together

Cadence

Weekly half-hour working sessions, a monthly budget-allocation review on blended MER, and a longer session before each drop to lock the prospecting guarantee.

Client side

The e-commerce manager owned campaign settings; the merchandiser owned drop dates; the finance contact attended allocation reviews and signed the MER numbers.

Decisions

Budget shares were settled at allocation reviews against pre-agreed rules; mid-cycle changes required the merchandiser's drop calendar as justification.

They provided

The drop calendar a full quarter out, promo-market lists for the holdout design, and order-data access so blended MER could be computed honestly.

What changed

The headline: retargeting's platform-reported roas versus its measured incremental roas — the gap, stated plainly, re-priced the line — 6.2× → 1.4×, read from Geo-holdout test results. A second check: blended mer over the quarter after re-balancing at +18%.

Drop launches stopped being budget negotiations. The merchandiser plans the calendar knowing the prospecting line will be there; the e-commerce manager stopped babysitting Advantage+ daily. The retargeting line is discussed with its measured contribution attached, which ended a long habit of quoting flattering numbers in team meetings. Allocation reviews run on one number everyone can verify, and the quarterly conversation moved from 'why did ROAS fall' to 'what did we buy' — a different question with better meetings.

The result was read from Geo-holdout test results against the pre-engagement baseline over the stated window, with a guardrail check on blended mer over the quarter after re-balancing. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.

What they own now

• The budget-share model — prospecting, retargeting, Advantage+ quotas with the exclusion map.
• The holdout test design and its results, reusable for any line's incrementality question.
• The blended-MER allocation template for quarterly reviews.
• The drop-calendar briefing template that locks prospecting guarantees per launch.
• Written rules for when Advantage+ budget can and cannot be touched mid-cycle.