The Pre-Launch Audit: Catching Feed and Asset Group Problems Before They Cost You
Objective: Given a supplied product feed export and an asset group screenshot for a new Performance Max campaign, catch the two failure patterns the lesson warns about before launch: feed hygiene gaps and undifferentiated creative across audiences.
You're the paid media analyst at Nykaa (FSN E-Commerce Ventures) reviewing a new PMax campaign for a skincare sub-brand before it goes live.
Check the feed for the specific hygiene gaps the lesson names, missing GTINs and thin titles, then check whether the asset groups are actually differentiated by audience or just copy-pasted.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, required for any Shopping/PMax feed
Free, no account friction
The process
2 steps
Step 01 of 02
The lesson's Common Mistakes callout names turning on PMax before the product feed is cleaned, missing GTINs, thin titles, then blaming the algorithm for weak results.
The export has 40 SKUs. 6 rows have a blank GTIN column, and 11 titles read as just the product name with no attributes ('Vitamin C Serum' instead of 'Nykaa Vitamin C Serum 30ml, Brightening'). Is this feed launch-ready?
Procedure
- Filter the export for blank GTIN values, count them (6 of 40)
- Filter for titles under 5 words with no attribute (11 of 40)
- Flag both sets as blockers, not warnings, before enabling the campaign
- Draft the required title fix pattern: Brand + Product + Size + Key Attribute
Feed hygiene check, 40 SKUs Missing GTIN: 6 rows (15%) Thin titles (<5 words, no attribute): 11 rows (27.5%) Launch-ready: NO — 17 of 40 SKUs (42.5%) need fixes first
Healthy
0 missing GTINs, every title follows Brand + Product + Size + Attribute, feed passes Merchant Center diagnostics with zero item-level disapprovals.
Unhealthy
Launching with 42.5% of SKUs carrying a hygiene gap, then reading a weak PMax result three weeks later as an algorithm problem instead of a feed problem.
What this means
PMax can only advertise what the feed accurately describes. A gap in GTIN or title data doesn't just risk disapproval, it starves the algorithm of the attributes it needs to match the right shopper.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| New PMax campaign underperforms in week 1 | Check feed diagnostics before touching bids or budget | 30 min |
Step 02 of 02
The lesson states creative has replaced manual targeting, the algorithm learns who to reach from who engages with each asset group, and warns against uploading the same three assets to every asset group.
This campaign has 3 asset groups (Skincare Starters, Anti-Aging, Gift Sets) but all three use the identical 5 product images and 1 generic headline. Is this set up correctly?
Procedure
- List all assets currently assigned to each of the 3 asset groups
- Mark any asset reused across 2+ groups
- Identify which group needs UGC-style creative (Skincare Starters, price-sensitive) vs. which needs before/after proof (Anti-Aging)
- Recommend at least 2 unique assets per group before launch
Asset group audit Skincare Starters: 5 images, 1 headline — all 5 shared with the other 2 groups Anti-Aging: same 5 images, same headline Gift Sets: same 5 images, same headline Unique assets per group: 0 of 3 groups pass
Healthy
Each asset group has at least 2 unique images and a headline written for that segment's motivation.
Unhealthy
Three asset groups that are functionally one asset group with three names, the algorithm has no differentiated signal to learn from.
What this means
Naming asset groups differently does nothing if the assets inside them are identical. Differentiation has to live in the actual creative, not the label.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| PMax asset groups all show similar audience overlap in reporting | Audit for shared assets across groups and replace duplicates | 30 min |
Final deliverable
A pre-launch audit memo listing every feed hygiene gap and every asset group differentiation gap, with fixes required before the campaign goes live.
See a reference example
Allbirds, PMax pre-launch audit (excerpt) Feed: 2 of 35 SKUs missing GTIN (fixed before launch) Asset groups: Everyday Wool vs. Trail Runners now use distinct hero images and segment-specific headlines Status: launch-ready
Success criteria
You're done when you can:
- Correctly counts and flags every missing-GTIN and thin-title row
- Correctly identifies that all 3 asset groups share identical creative