HowtoStructureGoogleShoppingCampaignsbyMargin-APracticalGuide

Most Google Shopping accounts are structured by campaign convenience - one campaign per channel type, products grouped by category or brand, a single ROAS target applied to everything. This structure is easy to manage. It is not designed to generate profit. Here is the alternative - a margin-informed campaign architecture that allocates budget and sets targets based on what each product actually contributes to your business.

Why structure matters more than bidding

The most common approach to Google Shopping optimisation focuses on the bid strategy: which bidding algorithm to use, what target to set, how aggressively to respond to auction signals. These decisions matter. But they matter significantly less than the structure within which they operate.

A bid strategy is a mechanism for allocating budget toward conversions. It operates within the structure you create. If the structure groups high-margin and low-margin products together under a single conversion target, the bid strategy will optimise toward the target - indiscriminately, across products with fundamentally different commercial profiles.

A 40% margin product and a 15% margin product in the same campaign with the same ROAS target will receive similar bids for similar searches. The 40% margin product can sustain a CPA of £28 before the campaign breaks even (at 3.5x ROAS and £98 AOV). The 15% margin product breaks even at a CPA of £10.50 on the same order value. The same bid optimisation treats them identically. One is generating profit. The other is destroying it. The ROAS looks fine.

Structure is the intervention that separates these two products and allows the bid strategy to do the right thing for each of them. Without it, the bid strategy cannot tell the difference.

Step one: Build your margin tiers

Before any campaign restructure, you need to segment your product catalogue by margin. This is the commercial input that everything else depends on.

Pull your COGS data by product or product category from your P&L or Shopify cost data. Calculate gross margin as (Revenue − COGS) ÷ Revenue for each product or category. Where you have return rate data by product, factor it in: effective margin = (Revenue − COGS) × (1 − return rate) ÷ Revenue.

From this data, build four tiers:

• Tier 1 - High margin (35%+): Economics strongly support aggressive Google Ads investment. These products can sustain higher CPAs and benefit from aggressive bidding to maximise volume within their profitable range.
• Tier 2 - Mid margin (20-35%): Google Ads can be profitable with disciplined targeting. Tighter bid management and more frequent optimisation - the margin of error is smaller.
• Tier 3 - Low margin (under 20%): Difficult to run profitably. Very tight CPA targets and hard floors. In many cases, exclude from paid shopping entirely and let them convert through organic and direct.
• Tier 4 - Exclude: Cannot be advertised profitably at any realistic CPA. Heavy clearance lines, loss-leaders, items with very high return rates that compress effective margin below any sustainable threshold.

This segmentation does not need to be SKU-precise to start. Tier assignment at product category level - dresses, outerwear, footwear, accessories - is sufficient to begin a margin-informed structure and can be refined to SKU level as you develop the data infrastructure to support it.

Step two: Apply custom labels to the feed

Custom labels are feed attributes that you define - Google does not use them for matching, but they flow through to your campaign structure and allow you to create product groupings based on any criteria you choose.

In your feed management tool or directly in the supplemental feed, add a custom label column - typically custom_label_0 for margin tier. Assign each product the value that corresponds to its margin tier: tier1-high, tier2-mid, tier3-low, tier4-exclude. Products with tier4-exclude can be excluded from all Shopping campaigns using a product filter.

If you are also segmenting by new product status, bestseller designation, or seasonal relevance, you can use additional custom label columns (custom_label_1 through custom_label_4) for these attributes. Each additional dimension gives you more precise control over how Google allocates budget across your catalogue.

Once custom labels are in the feed, they are available in your campaign structure as grouping criteria. This is the mechanical link between your commercial segmentation and your campaign architecture.

Step three: Build the campaign architecture

With margin tiers applied as custom labels, you can build a campaign structure that manages each tier separately with appropriate targets. For a standard Shopping setup (not Performance Max), the architecture looks like this:

• Campaign 1 - Tier 1 (High Margin): Tier 1 products only. ROAS target 1.5-2x break-even - aggressive, because the margin supports it. Budget uncapped within the profitable range. Target ROAS, with Maximise Conversion Value as a fallback if volume is insufficient for learning.
• Campaign 2 - Tier 2 (Mid Margin): ROAS target 1.1-1.3x break-even - profitable but tighter. Budget allocated proportionally to the margin opportunity. More frequent optimisation required.
• Campaign 3 - Tier 3 (Low Margin): If you advertise them at all. ROAS target at or slightly above break-even. Hard daily budget cap. Review monthly - products consistently above the CPA ceiling get demoted to Tier 4.

Tier 4 products are excluded at campaign level using a product filter, or in the supplemental feed using the excluded_destination attribute.

This structure gives Google's bidding algorithms clear, separate signals for each margin tier. The algorithm for Tier 1 is optimising toward a different target, on different products, with a different budget, than the algorithm for Tier 2. They do not interfere with each other. Each develops its own learning based on the products and targets appropriate to its tier.

Applying the same logic to Performance Max

Performance Max does not support the same degree of structural control as Standard Shopping campaigns. But the margin-informed principle can still be applied through a combination of asset group structure, listing group segmentation, and campaign-level product filters.

• PMax Campaign 1 - High Margin: Tier 1 custom labels only, applied via listing group filter. Asset groups structured by product category within Tier 1. Target ROAS set aggressively relative to Tier 1 break-even. Brand exclusions applied at campaign level.
• PMax Campaign 2 - Mid/Controlled: Tier 2 products with a tighter ROAS target. Separate campaign to prevent Tier 1's learning from blending with Tier 2's economics.

Tier 3 products are either excluded from PMax entirely or placed in a separate campaign with a very tight target and a low daily budget cap. Tier 4 products are excluded via listing group filter or feed-level exclusion.

This approach is more complex than running all products in a single PMax campaign, but it prevents the most expensive outcome of undifferentiated PMax management: the algorithm spending your budget on high-volume, low-margin products and starving your high-margin products of impression share because they have less historical conversion volume to draw on.

Setting the right targets for each tier

The target for each campaign tier should be derived from the break-even ROAS for that tier, with an aggressiveness multiplier that reflects your growth objectives and risk tolerance.

• Tier 1 target ROAS = Break-even ROAS (Tier 1) × 1.5-2.0
• Tier 2 target ROAS = Break-even ROAS (Tier 2) × 1.1-1.3
• Tier 3 target ROAS = Break-even ROAS (Tier 3) × 1.0-1.1

The multiplier for Tier 1 is high because the margin supports aggressive volume acquisition - you want to scale as much profitable revenue as the market allows within the efficient range. The multiplier for Tier 2 is lower because the margin ceiling is tighter. The multiplier for Tier 3 is close to 1.0 because you are at break-even - any lower and you are deliberately running a loss, which requires explicit commercial justification.

These targets should be reviewed when margin changes. Supplier cost increases, new promotional pricing, or changes in your product mix all affect the break-even thresholds that anchor your targets. An annual target review is insufficient. Quarterly at minimum; monthly for categories with volatile cost structures.

Monitoring and maintaining the structure

A margin-informed structure requires more active maintenance than a flat structure, because it has more components and more dependencies.

New products entering the catalogue need to be tier-assigned before they go into campaigns - not after. A product that enters Tier 1 campaigns without a custom label will be treated as a Tier 1 product regardless of its actual margin, potentially receiving aggressive investment on thin margins.

Products whose margin changes - through supplier cost increases, promotional pricing, or changes in return rate - need to be reassigned to the appropriate tier. A custom label value that was correct six months ago may be wrong today if the product's economics have changed.

Campaign performance at tier level should be reviewed monthly against the break-even threshold for that tier. The question is not "is this campaign hitting its ROAS target" but "is this campaign generating positive contribution margin at its current CPA?" These are related but not identical questions. A campaign can hit a ROAS target that was set incorrectly and still be running below break-even.

The monthly review question for each tier

For each campaign tier, calculate:

Actual POAS = (Campaign revenue × Tier gross margin) ÷ Campaign spend

• If POAS is above 1.0: the campaign is generating gross profit on its spend. The question is whether there is room to scale further within the profitable range.
• If POAS is below 1.0: the campaign is destroying margin. Either the target needs to be raised, the product mix within the tier needs to be reviewed, or the tier assignment needs to be reconsidered.

The transition from a flat to a tiered structure

Moving from an existing flat campaign structure to a tiered, margin-informed architecture takes time and carries transition risk. The algorithm needs to rebuild its learning in the new structure. There will typically be a period of 4-6 weeks during which performance is less stable than in the established flat structure.

The recommended approach is to transition incrementally rather than rebuilding everything simultaneously. Start with Tier 1 - your highest-margin products. Create the Tier 1 campaign, populate it with your high-margin products via custom label filter, and run it alongside the existing flat structure for 4 weeks. Once it has sufficient learning and stable performance, begin transitioning Tier 2. Tier 3 last, with careful monitoring of its CPA ceiling.

This incremental approach limits the disruption to your highest-spend period and gives each campaign tier sufficient time to learn before further structural changes complicate the signal.

The result, after the transition period, is an account that manages your advertising investment the way a commercially sophisticated business manages any investment: with different risk tolerances, different return requirements, and different budget allocations for assets with different economic profiles. Not a single blended target that treats everything as equivalent regardless of what it actually contributes.