The Labelizer Playbook: Tier Your Feed by Margin, Watch ROAS Jump

A jewelry brand was running 2,400 SKUs in one undifferentiated Shopping campaign. Their hero bestsellers got the same bid as their long-tail filler. One Target ROAS across the board. Seven days after tiering the feed, ROAS was up 38%. Here is exactly what they did.

Why one campaign is a trap

Most brands start Shopping with one campaign, one product group, one Target ROAS. It is fine at 50 SKUs. It collapses at 2,000.

The problem: Google bids on expected margin, but does not know your actual margin. It bids on expected conversion velocity, but not your actual velocity. Without labels, every SKU looks identical to the auction, and your budget flows to whatever converts fastest regardless of whether you make money on it.

The three tiers that matter

Hero SKUs: high margin (60%+ gross) and high 30-day velocity. These are the products where every extra click is likely profit. Bid aggressively.

Workhorses: mid margin (30-60%) or mid velocity. Steady state products, the body of the catalog. Bid for a healthy but not heroic ROAS.

Bleed SKUs: low margin (under 30%) or low velocity (fewer than 3 sales in 30 days). Cap their spend or exclude them.

Those three buckets are enough to win. You can split further later.

How to label without Scaley

Export your full product catalog with margin and 30-day conversion count. In Google Sheets, write a tier formula: if margin >= 60% and velocity >= 10 sales, tag "hero". If margin < 30% OR velocity < 3, tag "bleed". Everything else is "workhorse".

Push the tier back to Merchant Center as a custom label (custom_label_0 works). Rebuild your Shopping campaigns around those labels: one campaign per tier, different Target ROAS each.

Do this manually for a one-shot restructure. Do it daily if you want the tiers to stay current. That is what the labelizer automates.

The Seven-Day ROAS Lift, Explained

When the jewelry brand restructured into three tiers with per-tier Target ROAS (4.0, 3.0, 2.2), two things happened fast.

Hero SKUs got more budget. Google stopped starving the 60%-margin bestsellers to fund a low-margin product that converted at a similar rate.

Bleed SKUs stopped eating budget. Products that were losing money on every sale got capped, freeing up spend for the middle.

Within a week, blended ROAS was up 38%, hero SKUs were carrying more volume, and the team did not have to touch a single bid manually.

What to do next

If your Shopping campaigns are not tiered by margin and velocity, that is usually the single highest-ROI change in the account. The only real decision is whether to do it once manually or run it daily with a tool like the labelizer.

Either way: stop treating your catalog as one undifferentiated blob. Your top 10% of SKUs deserve a different strategy than your bottom 10%.