A clean product feed first,
then Performance Max can actually learn
Performance Max is only as good as the feed and assets behind it. We have rebuilt feeds with a third of products broken by wrong prices or dead links before a single campaign goes live, because no amount of bidding skill fixes a bad feed.
Where the money leaks today
Performance Max hides its mechanics behind automation, which means a broken feed or a missing exclusion list burns budget quietly instead of obviously. We have seen feeds where a third of products were disapproved for missing identifiers, where prices in the feed did not match the site, and where the campaign spent against branded search terms that would have converted for free organically.
Because Performance Max optimises across Search, Display, YouTube and Gmail at once, a mistake in the feed or the asset group does not just hurt one channel, it hurts the whole campaign’s ability to learn where your actual buyers are.
Because Performance Max automates so much of the targeting decision, the feed and the signals you feed it are the main levers left to pull; a well-run account spends most of its effort there rather than fighting the automation. This is also why we treat the first few weeks as a deliberate learning phase rather than a performance window, resisting the urge to change settings daily in response to normal early volatility.
What we do
- Feed audit. Every disapproval, missing identifier and price mismatch found and fixed, because the feed is the product the algorithm is actually selling.
- Asset groups built with intent. Grouped by category, margin or new-collection status, with images, headlines and descriptions written for the format, not pasted from a product page.
- Exclusions that protect what already works. Branded terms and existing customer lists excluded or treated separately so Performance Max spends on finding new buyers, not rebuying your own brand searches.
- Signals fed deliberately. Customer match lists and conversion value rules set so the algorithm knows which customers and which conversions are worth more.
- Reporting split by value. New-customer acquisition tracked separately from repeat purchase, so you can judge the campaign on the metric that matters to your business.
What we need from you
- Access to Google Merchant Center and the Google Ads account.
- A product feed we can audit, either your platform’s native feed or a custom one.
- Clarity on your margin by product or category, if you want asset groups split by profitability.
- A customer list export if you want new-customer bidding, following Google’s data handling rules.
- Patience through the learning phase: Performance Max needs weeks, not days, to find its footing.
How we measure
We report new-customer value separately from returning-customer value, since Performance Max’s own dashboard blends both unless you set it up correctly. Feed health is checked weekly: disapprovals, out-of-stock items showing as available, price mismatches, all the quiet failures that do not show up as a single alarming number but add up.
Monthly reporting shows ROAS by asset group alongside the feed health score, so a dip in performance can be traced to a feed issue instead of guessed at.
We also watch for search term insights the campaign surfaces, since they reveal which queries Performance Max is actually matching your products against, useful for catching mismatches early. Where the mismatch traces back to a weak product title or missing attribute, we fix the feed directly rather than trying to compensate for it through exclusions alone.
Price and timeline
| Option | Price | What’s included | Timeline |
|---|---|---|---|
| Launch or audit | from $1,000 | Feed audit and fixes, Merchant Center cleanup, first asset groups live. | 2 weeks |
| Monthly management | from $1,200 / month | Weekly feed health checks, asset group iteration, monthly new-customer reporting. | monthly, no lock-in |
| Full control, handover to your team | from $2,800 | Feed pipeline, exclusions and signals documented and handed to your team or developer. | 4 to 6 weeks |
Related
This page sits under our Performance marketing and E-commerce work. See also Ecommerce catalog ad campaigns, Conversion tracking and capi setup, Media mix and budget planning for adjacent paid-channel services.
For the automation side of reporting and budget decisions, see ad performance reporting. For real numbers behind these benchmarks, read the Thailand D2C store audit and rebuild case study and sports-nutrition sales x2.7 case study.
Ready to see what this would look like for your account? Get in touch and we will look at your current setup in the first call.
FAQ
How much does this cost?
Launch starts from $1,000 for the feed audit and first campaigns; monthly management starts from $1,200, depending on catalogue size and how many markets you sell in.
How long until we see results?
Launch takes about 2 weeks once the feed is clean. Performance Max typically needs 4 to 6 weeks to leave its automated learning phase and show a stable read, which we explain up front so early weeks are not misread.
What ad budget do we need alongside this?
Feed-based campaigns need enough conversion volume to let the algorithm learn, which usually means at least $1,000 to $1,500 a month for a single market catalogue of meaningful size.
What do you need from us?
Merchant Center and Ads access, a feed we can audit, and margin or category data if you want asset groups split by profitability rather than treated as one undifferentiated catalogue.
How do you report on performance?
Monthly ROAS by asset group, feed health checks weekly, and new-customer value reported separately from repeat purchase so the number reflects what the campaign is actually doing for your business.