E-commerce PPC measured in profit,
not return on ad spend.

The core problem

Return on ad spend doesn't mean much without margin

Most e-commerce accounts optimize toward ROAS because it's the number Google puts on the screen. On its own, it doesn't tell you whether you're making money.

Take a $100 product with 30% contribution margin, so $30 before advertising. At a 4x return you spent $25 to sell it and kept $5. The dashboard shows a healthy 4x. Now scale it ten times: you've added ten times the volume, returns and support load for a few hundred dollars. The same 4x on a 70% margin product is a genuinely good business.

ROAS targets should come from margin, product by product. A single account-wide target means overspending on your thin-margin products and underspending on your best ones. Working out real per-product contribution margin, including shipping and returns, is where we start.

Two other numbers belong in the same conversation. Average order value sets how much room you have, since a $40 AOV store and a $400 AOV store can't run the same bids. And blended ROAS, total revenue divided by total ad spend across every channel, is usually closer to the truth than the in-platform number, because Google and Meta will each take credit for the same order. If you're chasing first-time buyers, it's worth tracking new customer acquisition cost separately.

The product feed is your targeting

In Search, you bid on keywords. In Shopping and Performance Max there are no keywords. Google reads your product feed to decide which searches you show up for.

That makes feed work the highest-impact job in most Shopping accounts, and it's usually the most neglected, because it looks like data entry. What matters:

• Titles that lead with what people search for. Brand, product type, and the attributes buyers actually type: size, color, material, model. Not your internal SKU naming.
• Product type and Google product category filled in properly, since they affect which auctions you enter.
• GTINs and brand, which affect eligibility and matching.
• Supplemental feeds to correct and enrich fields without rebuilding the source.
• Custom labels carrying margin band, season or stock level, so campaigns can bid differently on products that deserve different treatment.

Merchant Center disapprovals

A disapproved product doesn't show, and a suspended account shows nothing at all. The usual causes are straightforward: price or availability mismatches between the feed and the landing page, missing or inconsistent GTINs, image policy issues, and incomplete shipping or returns settings. Any of them can quietly pull part of your catalog out of the auction while spend continues on what's left.

Shopping and Performance Max

Performance Max takes a budget and a goal and allocates on its own. That works well when the inputs are right and hides problems when they aren't. Two decisions do most of the work:

• Segmentation. Splitting by margin band, best sellers versus the long tail, or new versus returning customers gives you control. One campaign holding the entire catalog gives you a single lever.
• Asset quality. Thin assets mean Google fills the gaps itself, often on placements you wouldn't have chosen.

Standard Shopping still earns its place where control matters more than reach, particularly for a small, high-value catalog.

What we fix first

1. Conversion tracking and revenue accuracy. If reported revenue doesn't match your store's own numbers, nothing after this point is reliable.
2. Margin by product. Real contribution margin including shipping and returns, so targets come from math rather than habit.
3. Feed quality and Merchant Center health. Get the whole catalog eligible and well described before paying to promote it.
4. Campaign structure. Segmented so budget can move toward the products that can carry it.
5. Then scale the products the first four steps proved are worth scaling.

Most of the accounts we audit started at step five.