Google Ads Shopping study materials
Google Ads Shopping study resources

A 4.5x ROAS Channel That May Be Losing Money

Google Ads Shopping · Put your knowledge to work in context.

All case studies

A 4.5x ROAS Channel That May Be Losing Money

An outdoor-gear retailer runs one standard Shopping campaign with tROAS 400% and a $12,000 monthly budget. Last 30 days by category: Tents - revenue $90,000, gross margin 8%, ROAS 620%; Packs - revenue $25,000, margin 42%, ROAS 280%; Stoves - revenue $60,000, margin 15%, ROAS 400%. Spend allocation is left to the bid strategy, and Packs receives only 6% of campaign spend while Tents takes 58%. Finance reports the channel is 'profitable at 4.5x ROAS' and wants to raise the budget. The feed uses custom_label_0 for season only; custom_label_1 and custom_label_2 are empty. No cost data exists in Merchant Center or the conversion tag. Analytics can join product cost to SKU in the warehouse but cannot pass it to Google Ads.

1. Test finance's claim that the channel is profitable at the current ROAS, showing the arithmetic.

Read the worked answer

Spend by category = revenue ÷ ROAS:

  • Tents: 90,000 ÷ 6.20 = $14,516
  • Packs: 25,000 ÷ 2.80 = $8,929
  • Stoves: 60,000 ÷ 4.00 = $15,000
  • Total spend $38,445; total revenue $175,000 → blended ROAS 4.55x, matching finance's figure.

Gross profit by category = revenue × margin:

  • Tents: $7,200
  • Packs: $10,500
  • Stoves: $9,000
  • Total $26,700

Profit after ad spend = 26,700 − 38,445 = −$11,745.

The channel is running at a loss. The error is treating ROAS as a profitability metric: ROAS only becomes profitable above the break-even point set by margin, and break-even ROAS = 1 ÷ margin.

  • Tents: 1 ÷ 0.08 = 1,250%
  • Packs: 1 ÷ 0.42 = 238%
  • Stoves: 1 ÷ 0.15 = 667%

So Tents at 620% and Stoves at 400% are both below break-even even though both look healthy against a 400% target. Only Packs, at 280% against a 238% break-even, is making money. A single campaign-level tROAS of 400% is structurally unable to fix this.

2. Design a margin-tier structure using the feed and campaign settings, including target ROAS values for each tier. Explain why conversion value rules are not the right tool here.

Read the worked answer

Tier structure

  1. Populate custom_label_1 from the warehouse join (SKU → gross margin band), uploaded as a supplemental feed keyed on SKU. Bands, not raw margin: margin_lt10, margin_10_24, margin_25plus. Keep the mapping in one maintained table.
  2. Create one campaign per band, since target ROAS is a campaign-level (or portfolio-level) setting and cannot be applied per product group within a campaign. Same feed, different product groups defined by custom_label_1.
  3. Set targets above break-even with a buffer for the target's own error:
  • Tents (margin_lt10): break-even 1,250% → either bid at 1,300%+ and accept low volume, or exclude the category and treat it as a merchandising decision, not a bidding one. Do not let it inherit the 400% target.
  • Stoves (margin_10_24): break-even 667% → target 700–750%.
  • Packs (margin_25plus): break-even 238% → target 300–320%; this is the tier that should absorb budget, and it currently gets 6% of spend.
  1. Reallocate budget toward Packs only after the tiers have run long enough to stabilize; expect Tents volume to fall sharply and be explicit with finance that this is the intended outcome.

Why not conversion value rules: they adjust conversion value by conditions such as audience, location or device. They cannot key off product or SKU, so they cannot express per-category margin. Feed labels plus per-campaign targets are the mechanism that operates at product level.