Google Shopping product groups are the structural layer that determines how Google allocates your budget across your catalog. A well-built product group hierarchy is the foundation of margin-based bidding: it lets you assign different bids or target ROAS values to products based on what they actually earn you, not what Google's automation decides to prioritize. Without deliberate product group segmentation, your highest-margin products compete for budget against low-margin SKUs in a single pool, and the algorithm optimizes for volume, not profit.
This six-step framework walks you through auditing your current structure, choosing the right segmentation hierarchy, labeling products by margin tier, setting bids that reflect real economics, connecting everything to Performance Max asset groups, and building a maintenance routine that keeps performance from decaying. By the end, you will have a product group architecture that gives you real bidding control at the SKU level.
You will need: access to Google Ads and Google Merchant Center, your product feed (ideally with margin or COGS data available), and either a spreadsheet or feed management tool for custom label assignment.
Before You Start: What You Need In Place
Before restructuring product groups, confirm three things. First, your Merchant Center feed must be healthy. Disapprovals, missing GTINs, or stale pricing undermine any structure you build on top. If feed quality is a concern, read how an ecommerce brand fixed Google Shopping feed quality and doubled new customer growth before proceeding. Second, you need margin data. This does not have to be perfect, but you need at least a rough tier system (high, medium, low margin) for your SKUs. Third, you should have at least 30 days of Shopping campaign data so the audit step provides meaningful signal. If you are launching fresh, start with the segmentation hierarchy and come back to the audit once data accumulates.
Step 1. Audit Your Current Product Group Structure Before Touching Anything
The first action is diagnosis, not surgery. Open your existing Shopping or Performance Max campaign, navigate to the product groups tab, and document exactly how products are currently grouped. Most accounts have one of two problems: everything lives under a single "All Products" group with one bid, or there is a half-built hierarchy that someone started and never finished.
How To Pull A Search Terms Report By Product
In Standard Shopping, go to the Products tab, select individual product groups, and then view the search terms report filtered to that group. In Performance Max, this is harder because Google restricts search term visibility, but you can still go to Insights and then Search Terms to see category-level query data. Export this data. What you are looking for is which search queries are triggering which products, and whether high-cost queries are attached to low-value SKUs.
Identifying Which Products Are Eating Budget Without Converting
Sort your product-level data by cost descending. Flag any product that has spent more than twice your average cost per conversion without generating a single sale. These products are not just underperforming; they are actively stealing budget from products that would convert. In many accounts, the bottom 20% of SKUs by conversion rate consume 30-40% of total Shopping spend. This is the waste your new structure will eliminate.
If you have already seen the damage a flat product group structure causes, this case study on rebuilding product groups and recovering ROAS in 60 days shows what a proper restructure looks like in practice.
Step 2. Choose Your Segmentation Hierarchy
Your segmentation hierarchy is the order of subdivision within your product groups. Google lets you segment by category, brand, product type, custom label, item ID, condition, and channel. The hierarchy you choose determines how granular your bidding control can be.
Segmenting By Category Versus Brand Versus Product ID
Category-first works well for retailers with diverse product lines where margin profiles differ across departments (apparel vs. accessories vs. electronics). Brand-first is better when you carry multiple brands and margin varies more by brand than by category. Product ID gives maximum granularity but becomes unmanageable above a few hundred SKUs.
When Custom Labels Beat Native Google Attributes
Custom labels are the most powerful segmentation tool because they let you encode business data that Google does not natively understand: margin tier, seasonality, promotional status, inventory velocity, or competitive positioning. You get five custom labels (0 through 4) in Merchant Center. Use at least one for margin tier. The others can encode things like "hero product," "clearance," or "new arrival." Custom labels beat native attributes whenever the bidding distinction you need does not map to a standard Google taxonomy.
The Right Depth For Your Catalog Size
Do not over-segment. For catalogs under 500 SKUs, two levels of hierarchy (e.g., custom label for margin tier, then product type) usually suffice. For 500 to 5,000 SKUs, three levels give enough control without drowning in micro-groups. Above 5,000 SKUs, use custom labels aggressively and lean on feed rules to automate assignment rather than trying to manually classify everything. The goal is groups large enough for Google's algorithms to gather statistical signal, but specific enough that you are not bidding the same amount on a $200-margin product and a $5-margin product.
Step 3. Build Product Groups Around Margin And Intent
This is where the actual restructuring happens. You are going to create product groups that reflect what each product is worth to your business, not just what category it falls into.
Assigning Custom Labels For Margin Tiers
In your feed, create a custom label (custom_label_0 is the standard choice) with values like "high_margin," "mid_margin," and "low_margin." The thresholds depend on your business, but a common split is: high margin is anything above 50% gross margin, mid margin is 25-50%, and low margin is below 25%. If you do not have exact margin data, use average selling price minus landed COGS as a proxy. Assign these labels in Merchant Center via feed rules or through your feed management tool, then subdivide your product groups by this custom label first.
Separating Hero Products From Long-Tail Inventory
Hero products are SKUs that drive disproportionate revenue or serve as entry points to higher-LTV customer journeys. These get their own product groups, ideally at the item ID level, so you can set aggressive bids independently. Long-tail inventory (products that sell infrequently but contribute to catalog completeness) gets grouped together at a higher level with conservative bids. The principle: your best products deserve individual attention, and everything else gets managed as a cohort.
Creating An Everything Else Group With A Deliberate Low Bid
Every product group hierarchy should end with an "Everything else" group. This catches any SKU you have not explicitly segmented. Set this group's bid deliberately low, not zero. A low bid (or a high tROAS target) ensures these products still serve if there is a strong signal, but they do not consume meaningful budget. Treat this group as a monitoring lane: if a product starts generating conversions from the "Everything else" bucket, that is your signal to promote it into a proper segment.
Step 4. Set Bids That Reflect Actual Product Economics
Bids should map to what you can afford to pay for a click given the product's margin and expected conversion rate. This is where most accounts fail: they use a single target ROAS across the entire campaign, which forces high-margin and low-margin products into the same economic frame.
tROAS By Product Group Versus Blended Account ROAS
A blended account-level ROAS target is a blunt instrument. If your high-margin products can profitably run at 300% ROAS and your low-margin products need 800% ROAS to break even, a blended 500% target under-bids on the former and over-bids on the latter. Set tROAS at the product group level. In Standard Shopping, this means separate campaigns or ad groups per margin tier with distinct bidding strategies. In Performance Max, this means separate asset groups or separate campaigns per tier.
For a deeper dive on choosing between tCPA and tROAS and setting targets that reflect actual business math, see how to set target CPA and target ROAS in Google Ads.
Max CPC Floors For Low-Volume Products
Low-volume products often lack the conversion data for Smart Bidding to work effectively. For these SKUs, consider running them in a Standard Shopping campaign with manual CPC or enhanced CPC rather than tROAS. Set a max CPC that reflects the margin: if a product has a $40 margin and a 2% conversion rate, your breakeven CPC is $0.80. Bid below that to maintain margin. This approach sounds old-school, but smart bidding can hurt conversion volume when data is thin, and manual control on low-volume SKUs is often the more profitable path.
Step 5. Connect Product Groups To Asset Groups In Performance Max
Performance Max does not use traditional product groups in the same way Standard Shopping does, but the segmentation logic still applies. Your product group structure needs to translate into asset group structure.
One Asset Group Per Product Segment: Why It Matters
Each asset group in Performance Max can target a specific set of products via listing group filters. Create one asset group per margin tier or per product segment. This gives Google's algorithm a clear signal about which products belong together and lets you attach different creative assets (headlines, images, videos) that match the products in each group. A single asset group containing your entire catalog is the Performance Max equivalent of the "All Products" problem in Standard Shopping: zero control over where budget goes.
If you are weighing whether to run Performance Max, Standard Shopping, or both, this analysis of Performance Max versus Search campaigns breaks down when each structure wins.
Feed Filters That Enforce The Segmentation
Inside each asset group, use listing group filters tied to the same custom labels you set in Step 3. This is critical: if your feed labels do not match your asset group filters, products will leak into the wrong groups or get excluded entirely. Verify the filters after setup by checking the "Products" tab within each asset group to confirm the right SKUs appear. Any mismatch between your Merchant Center labels and your campaign filters will silently undermine the entire structure.
Step 6. Monitor And Prune On A Cadence
A product group structure is not a set-and-forget asset. Catalogs change, margins shift, and Google's algorithm adapts its behavior over time. You need a recurring review process.
The Weekly Product Group Review Routine
Every week, pull product-level performance for the past 7 and 30 days. Check three things: (1) Are any products in high-margin groups underperforming their ROAS targets? (2) Are any products in the "Everything else" group generating meaningful conversions? (3) Has any product's margin changed due to price changes or supplier cost increases? Update custom labels, adjust bids, and promote or demote products as needed. This review takes 20 to 45 minutes per campaign, depending on catalog size.
When To Merge Versus Split Product Groups
Merge product groups when individual groups have fewer than 10 conversions per month. Below that threshold, Smart Bidding lacks the data to optimize effectively, and you are better off pooling similar products together. Split product groups when a group contains products with meaningfully different margins or conversion rates. The decision is always the same: does this group contain products that should have different bids? If yes, split. If a group is too small for the algorithm, merge up.
Common Mistakes That Undo A Good Structure
Setting it up once and never revisiting. Catalogs are dynamic. New products enter without labels, prices change, and seasonal shifts alter what converts. A structure that is not maintained decays within weeks.
Over-segmenting into micro-groups. If a product group contains two SKUs that together get three clicks a week, Google has nothing to optimize against. You have precision without signal, which is worse than a blended group with enough data to bid intelligently.
Using the same tROAS across all product groups. This negates the entire purpose of the segmentation. If every group shares a target, you have the appearance of structure without the economic differentiation that makes it valuable.
Ignoring the "Everything else" group. Leaving it at the default bid means new products enter your catalog and immediately compete for budget at the same level as your hero products. Always set a deliberate, conservative bid on this catch-all.
Not connecting feed labels to campaign filters. If your Merchant Center custom labels say one thing and your campaign listing groups filter on something else, products end up in the wrong segments. Audit the connection after every feed update.
Rebuilding from scratch every quarter. Iterative adjustments beat full restructures. Every time you reset a campaign, you wipe out the learning data that Smart Bidding depends on. Exiting the learning phase faster matters, and constant restructures keep you permanently in it.
How groas Handles This For You
The six steps above work. They also require ongoing attention: weekly reviews, feed label maintenance, bid adjustments, Performance Max asset group management, and the judgment to know when to merge versus split. For agencies managing dozens of client catalogs, this is where execution bottlenecks hit hardest. For in-house teams, it is one more recurring task competing with everything else on the roadmap.
groas approaches this differently depending on what you need. For agencies running client accounts, the groas engine (available through the DIY product) automates the heavy lifting of product group segmentation and margin-based bid management across unlimited client accounts. Your team stays in control, keeps the client relationship, and uses the engine to execute at a scale that manual work cannot match. Start your 7-day free trial to see how it works on a live account.
For ecommerce teams with someone in-house who knows Google Ads, groas DWY pairs the engine with a senior strategist who works alongside your team. The engine handles the 24/7 execution, including product group optimization and feed-level bid adjustments, while the strategist advises on structure and catches what the data alone cannot explain. Get started through self-serve checkout, or apply if you are managing larger spend.
For businesses that want Google Ads fully handled, groas DFY means a dedicated strategist owns your entire account end-to-end, including the product group architecture, feed quality, margin labeling, and ongoing pruning. Nothing to log into or manage. Apply to see if it is the right fit.
The core difference: a proprietary engine trained on over $500 billion in profitable ad spend runs the execution around the clock, and a real human strategist makes the decisions that algorithms cannot. The gap between manual product group management and what groas delivers shows up in the numbers within the first few weeks.
Bottom Line: Product Groups Are Your Bidding Architecture, Not Just Folders
Product groups are the structural foundation of every Google Shopping bid decision. Get the structure wrong, and no amount of bid optimization fixes the underlying problem: budget flowing to the wrong products. Get it right, and you give yourself (or your bidding algorithm) the granularity to allocate spend based on actual product economics.
The six steps are straightforward: audit, choose your hierarchy, build around margin and intent, set differentiated bids, connect to Performance Max asset groups, and maintain on a cadence. The hard part is not knowing what to do. It is doing it consistently, across every SKU, every week, as your catalog evolves. That is exactly the kind of work that groas was built to handle, whether your team drives the strategy or groas owns it entirely.
Frequently Asked Questions
What Are Google Shopping Product Groups And Why Do They Matter For Bidding?
Google Shopping product groups are subdivisions within a Shopping or Performance Max campaign that let you group products and assign different bids or bidding targets to each group. They determine how Google allocates your budget across your catalog. Without deliberate product group segmentation, all your SKUs compete in a single pool, and Google's algorithm optimizes for volume rather than profit. A well-structured product group hierarchy gives you control over which products get aggressive bids and which ones run conservatively, so your ad spend reflects actual product economics instead of a one-size-fits-all target.
How Many Custom Labels Can I Use In Google Merchant Center For Product Segmentation?
Google Merchant Center gives you five custom labels, numbered 0 through 4. Each custom label can hold values you define, such as margin tiers (high, mid, low), promotional status, seasonality, or product velocity. Best practice is to use at least one custom label for margin-based segmentation, which directly supports differentiated bidding in your campaigns. The remaining labels can encode other business-specific attributes that Google's native product taxonomy does not capture, giving you bidding flexibility that category or brand segmentation alone cannot provide.
What Is The Best Segmentation Hierarchy For A Large Ecommerce Catalog?
For catalogs above 5,000 SKUs, the most effective approach is leading with custom labels for margin tier, then subdividing by product type or brand depending on where margin variation is greatest. Use feed rules in Merchant Center to automate label assignment rather than classifying manually. Keep groups large enough to accumulate at least 10 conversions per month so Smart Bidding has sufficient data. For catalogs under 500 SKUs, two levels of hierarchy usually provide enough control without creating data-starved micro-groups.
Should I Use The Same Target ROAS Across All Product Groups?
No. Using a single target ROAS across all product groups defeats the purpose of segmenting in the first place. High-margin products can profitably run at lower ROAS targets (meaning higher bids), while low-margin products need much higher ROAS targets to break even. Set tROAS at the product group or campaign level so each segment reflects its actual economics. A blended account-level ROAS target under-bids your best products and over-bids your worst, which is the exact problem margin-based segmentation is designed to solve.
How Often Should I Review And Update My Product Group Structure?
Weekly reviews are the standard cadence. Each week, check product-level performance over the past 7 and 30 days. Look for underperforming products in high-margin groups, products in the "Everything else" group that are generating conversions, and any SKUs where margin has shifted. Update custom labels and bid targets as needed. This process typically takes 20 to 45 minutes per campaign. For teams that find this recurring workload unsustainable, groas handles product group maintenance continuously through a proprietary engine that monitors and adjusts around the clock.
How Do Product Groups Work Inside Performance Max Campaigns?
Performance Max uses listing group filters within asset groups instead of traditional product groups. The segmentation logic is the same: you create one asset group per product segment and use listing group filters tied to your Merchant Center custom labels to control which products appear in each group. This lets you attach different creative assets and bid targets per segment. Without this structure, Performance Max treats your entire catalog as one pool, giving you no control over budget allocation.
What Should I Do With Products That Have Very Low Sales Volume?
Low-volume products often lack enough conversion data for Smart Bidding to optimize effectively. Consider running these SKUs in a separate Standard Shopping campaign using manual CPC or enhanced CPC rather than tROAS. Calculate your breakeven CPC (margin multiplied by conversion rate) and bid below it. This gives you direct control over spend while the product accumulates data. If a low-volume product starts converting consistently, promote it to a Smart Bidding group with the appropriate margin label.
Can groas Handle Product Group Segmentation Automatically?
Yes. groas is built to handle exactly this kind of recurring, data-intensive execution. For agencies, the groas engine automates product group segmentation and margin-based bid management across unlimited client accounts through the DIY product, with a 7-day free trial to see it in action. For ecommerce teams, groas DWY pairs the engine with a senior strategist who advises on structure while the engine runs 24/7. For businesses that want it fully handled, groas DFY means a dedicated strategist owns the entire product group architecture, feed quality, and ongoing pruning end-to-end.
When Should I Merge Product Groups Versus Split Them?
Merge when a product group has fewer than 10 conversions per month. Below that threshold, Smart Bidding does not have enough signal to optimize, and you gain more from pooling similar products. Split when a group contains products with meaningfully different margins or conversion rates that warrant different bids. The question is always the same: do the products in this group deserve different bids? If yes, split. If the resulting groups would be too small for the algorithm to learn from, keep them together.
What Is The Biggest Mistake Advertisers Make With Google Shopping Product Groups?
The single most common mistake is leaving all products in a single "All Products" group with one bid. This means your highest-margin hero products compete for budget against low-margin or non-converting SKUs, and Google's algorithm decides where spend goes based on predicted click-through rate rather than your business economics. The fix is the margin-based segmentation framework described in this guide: label products by margin tier, create product groups around those labels, and set differentiated bids that reflect what each product actually earns you.
.png)
