Google Shopping looks like a paid media channel until a marketplace agency has to explain why the same SKU is profitable on one channel, out of stock on another and still receiving Performance Max budget. Then it stops being a campaign problem. It becomes a routing problem.
That distinction matters for agencies with five or more people managing clients across Amazon, Walmart, bol, Kaufland, TikTok Shop, Shopify and Google Merchant Center. The Google Ads specialist sees ROAS. The marketplace operator sees Buy Box pressure. The feed specialist sees Merchant Center warnings. The inventory owner sees 19 days of stock. The account manager hears the client asking why Google is scaling a product Amazon also needs for Prime Week. Everyone is looking at the same business from a different dashboard.
The named mistake I see is treating Google Shopping as a separate performance channel while marketplace operations absorbs the trade-offs. A team optimizes titles, fixes disapprovals, launches Performance Max, creates custom labels and celebrates a 4.2 ROAS. Two weeks later, Amazon rank drops because stock moved elsewhere, Walmart ads lose momentum, the client asks why a low-margin variant got so much spend, and the agency has to rebuild the story in a spreadsheet. The Google campaign was not necessarily wrong. It was allowed to move demand without a product-routing decision.
My stance: marketplace agencies need a Google Shopping product routing ledger. Not another feed audit. Not a prettier PMax report. A SKU-level decision layer that says which products are allowed to receive Google Shopping demand, which should be protected for marketplaces, which need feed repair first, and which should not be advertised anywhere until margin, stock or product data improves.
What the public advice gets right
The existing advice around Google Shopping is useful. Productsup explains the mechanics well: Shopping ads are built from product feed data, not keywords; product titles, categories, GTINs, availability and landing-page consistency influence eligibility and performance; stale data can expire products; discrepancies between feed and landing page can trigger disapprovals or even account risk. Their agency case studies are also practical. Webrepublic needed a central environment for product feed exports, scheduling and diagnostics across multiple clients. Peak Ace worked with an 18,000-product data set and eight price comparison engines, using feed rules, templates and tracking to improve visibility.
Channable’s agency positioning is similarly clear: agencies need automation for shopping feeds, Google Ads, product listings and marketplace connections. ChannelEngine frames marketplace management around content, inventory, pricing, orders, compliance and analytics. Rithum’s marketplace best practices remind sellers to keep descriptions fresh, stay competitive on price, monitor performance, advertise and expand carefully. MerchantSpring speaks directly to agencies that need reporting across many channels and brands.
All of that is true. But most of it still treats Google Shopping, marketplaces, feeds and reporting as adjacent workstreams. The agency problem is sharper: when Google can create demand for the same product that Amazon, Walmart or bol also needs, who decides where the next unit of stock and the next euro of budget should go?
The missing layer: product routing before campaign optimization
A product routing ledger is the bridge between feed quality and commercial permission. It does not replace Merchant Center, Google Ads, marketplace connectors or client reports. It decides what those systems are allowed to do.
At minimum, the ledger should hold eight fields for every advertised SKU:
- SKU contribution margin after marketplace fees, fulfilment, returns and ad allowance.
- Stock runway by channel, not just total stock.
- Channel role: acquisition, profit harvest, rank defense, clearance, retail readiness or test.
- Feed eligibility: approved, limited, disapproved, missing attributes or landing-page mismatch.
- Ad permission: scale, maintain, learn, pause or block.
- Client decision owner: who can approve budget, assortment and margin exceptions.
- Current demand conflict: where the same SKU is already being pushed.
- Expiry date for the decision, because last week’s stock and margin can become fiction quickly.
This sounds operational because it is. It is also where agency margin is protected. Without the ledger, senior people keep re-litigating the same questions: why did Google scale this SKU, why did Amazon run out of stock, why did Merchant Center approve products the client does not want to push, why did the report show ROAS but finance show weak profit?
Scenario 1: RhineTech Electronics and the dangerous 4.2 ROAS
Imagine RhineTech Electronics, a German consumer electronics brand managed by a 12-person marketplace agency. The client sells 2,400 SKUs across Amazon.de, Kaufland, Otto, bol and Shopify. Google Shopping and Performance Max have an €18,000 monthly budget. On paper, the account looks healthy: 4.2 ROAS, 62% impression share on priority products and only 3% disapproved items in Merchant Center.
The routing ledger changes the conversation. Of the 2,400 SKUs, 320 have contribution margin below 12% after marketplace fees, warranty reserve and average return cost. Another 190 have fewer than 14 days of stock in the German warehouse. Forty-eight hero accessories are profitable on Shopify but weak on Amazon because the Buy Box is unstable and the referral-fee plus FBA route is heavier. Google does not know any of that unless the agency tells it through labels, exclusions and budget structure.
The named mistake would be raising PMax budget because blended ROAS looks good. The ledger says something more precise: keep €9,500 on the 740 SKUs with more than 21 days of stock and 18%+ contribution margin, cap €4,000 on learning groups, exclude the 320 margin-risk SKUs, and hold €4,500 until Amazon stock for the top 25 products is safe again. That is not anti-Google. It is pro-profit.
FiveX fits naturally here because the agency can group products by margin, stock runway and channel role, then connect ad spend and revenue back to SKU-level profitability. The Google Ads specialist no longer has to ask finance for a margin file every Friday. The marketplace lead can see whether a Google Shopping win is creating a marketplace constraint before the client sees it in the next QBR.
Scenario 2: Maple Home Co. and the marketplace budget trade-off
Now take Maple Home Co., a US home goods brand with Amazon, Walmart, Shopify and Google Shopping. The agency manages $42,000 in monthly media: $18,000 Amazon Ads, $9,000 Walmart Connect, $12,000 Google Shopping and $3,000 Meta retargeting. One storage basket line becomes the hero. Google Shopping reports 4.6 ROAS at a $28 average order value. Amazon shows 23% ACOS. Walmart looks slower at 3.1 ROAS.
If the agency only reads platform reports, Google wins the argument. But the product routing ledger adds the boring numbers that decide profit. The Google route has a 14% return rate because shoppers misread dimensions. Amazon converts lower-margin bundles but produces only 6% returns. Walmart has slower volume but better shipping economics through WFS. After returns and fulfilment, the Google hero SKU lands at 7% contribution margin, Amazon at 13% and Walmart at 16%.
The agency does not need to kill Google Shopping. It needs to route it. Shift $10,500 of Google budget away from the high-return basket, keep $3,500 for query learning and free-listing visibility, move $5,000 into Walmart products with better contribution margin, and use the remaining $5,500 to test revised titles and images that reduce dimension confusion. The client sees a commercial decision, not a channel turf war.
This is where FiveX reporting matters. A normal dashboard can show ROAS by channel. An agency operating layer should show the next-dollar decision: which SKU deserves spend after fees, returns, stock and channel role are visible. That is a different conversation with the client, and frankly a much better one.
Scenario 3: NordPeak Outdoor and the disapproval queue
Feed issues are not all equal. NordPeak Outdoor has 680 SKUs and 76 Merchant Center disapprovals. A generic feed audit says, “Fix the 76 issues.” The agency team hears 76 tickets, each apparently urgent.
The routing ledger ranks them differently. Twelve disapproved SKUs are top-margin products with 45+ days of stock and active Amazon campaigns already proving demand. Fix those first. Twenty-three are low-margin clearance items that the client does not want to promote if Google approves them. Put them in a blocked label until pricing changes. Fifteen have missing GTINs but no stock runway. Wait. The remaining twenty-six have landing-page mismatches that also affect Meta catalog quality, so they become a feed-system fix rather than a Google-only task.
In this version, the agency does not sell “Merchant Center cleanup”. It sells profit-safe prioritization. That is a stronger retainer story.
How to build the ledger in a marketplace agency
Start with the decision you want to protect: no product should receive scalable Google Shopping demand unless it has acceptable margin, enough stock, clean feed status and a clear channel role. Then build the workflow backwards.
First, define the product groups. Do not group only by category. Add commercial groups: profit builders, rank defenders, clearance, stock-risk, feed-repair, test and blocked. In FiveX, agencies can use product groups and product-level profitability to make this structure visible across reporting and ad decisions.
Second, set budget permissions. A product with 25% contribution margin and 40 days of stock can scale differently from a product with 9% margin and 11 days of stock. Create simple permissions: scale, maintain, learn, repair, protect or block. The point is not bureaucracy. The point is preventing every platform from independently deciding that it deserves the same SKU.
Third, connect feed work to commercial impact. Merchant Center disapprovals, missing attributes and landing-page mismatches should be sorted by profit exposure, not by the order in which they appear. If a fix unlocks €6,200 of profitable demand, it deserves priority. If it unlocks a SKU the client should not push, it can wait.
Fourth, create an approval path for exceptions. Sometimes a low-margin SKU should still receive Google Shopping demand because it introduces customers to a high-repeat category. Sometimes stock-risk products should be promoted because the inbound shipment is confirmed. Fine. But those are exceptions, and exceptions need an owner, a reason and an expiry date.
Finally, report the ledger, not just the platform. Clients do not need another screenshot of Merchant Center diagnostics. They need to know which products are approved for demand, which are blocked, what profit is at risk, and what the agency is doing next. FiveX can bring marketplace performance, ad spend, SKU margin, product groups and inventory signals into one reporting layer so the client sees the operating decision behind the campaign.
The trade-off agencies should accept
The product routing ledger will slow down a few campaigns. Good. Some campaigns deserve to be slowed down.
Speed is valuable when the product is ready. It is expensive when the product has weak margin, fragile stock, messy feed data or unresolved channel conflict. The operator voice here is simple: do not let a clean Google Shopping setup outrun a messy marketplace business case.
If you manage marketplace clients, Google Shopping is not outside the marketplace operating room. It pulls from the same catalog, competes for the same stock, influences the same price expectations and changes the same client profit. Treat it as part of the product routing system, and the agency starts making better decisions: fewer budget arguments, cleaner client reporting, less emergency feed work and more spend going to SKUs that can actually turn demand into profit.
That is the real job. Not “more Google Shopping performance”. More profitable demand, routed to the right product, at the right moment, with proof the client can trust.