← catspring-tea

Last 28 Days — Google Ads

Google Ads account 509-139-6447. Last 28 days (Mar 31 - Apr 27) vs previous 28 days (Mar 3 - Mar 30). Single Performance Max campaign.

Spend dropped 13% (from $821 to $715), clicks fell 20%, and impressions dropped 29%. The headline conversion number is broken — Google is reporting 373 conversions this period vs 5,951 last period, which is clearly a tracking issue (a single PMax campaign with a $30 average order value cannot produce 5,951 purchases on $821 spend). Conversion value is the only number worth trusting: $1,398 vs $6,840, an 80% drop. That's the real story — actual revenue from Google Ads collapsed, even though spend was nearly flat.

Daily spend started high (Apr 2-4 averaging ~$56/day) then dropped sharply through mid-April (Apr 17-23 averaging $14/day) before recovering Apr 24-27 ($36/day). The mid-April pullback looks like budget pacing or a manual adjustment — confirms that whoever's running the account was actively managing it.

Other key metrics

Conversion value tells the real story. Last period the campaign returned $8.34 for every $1 spent. This period it returned $1.95 — barely above breakeven. Both clicks and impressions dropped, so the campaign isn't just ranking lower for the same money, it's literally showing less. That's typical Performance Max behavior when conversion data degrades: the bidder pulls back automatically.

Conversion tracking is broken

Reported conversions:

Neither implied AOV is realistic for a tea brand with a $20-40 average order. Two things are happening:

1. Google Ads is counting an event that fires on every page view or session, not just purchases. Likely a "view content" or "begin checkout" event mis-mapped as the primary conversion.
2. The previous period's 5,951 figure is even more obviously broken — that would imply 8 conversions per click, which is impossible.

This needs to be fixed in conversion tracking before any optimization decisions can be made. The PMax bidding algorithm is currently optimizing toward a junk signal.

Campaign breakdown

Single-campaign account. Everything is in this one Performance Max campaign. Worth considering a campaign split (pure Search for branded queries, Shopping for product feed) once tracking is fixed — PMax is hard to diagnose when something goes wrong because the channel mix is opaque.

Monthly trend

December 2025 was the peak ($7,575 in spend) — likely a holiday push. January spent another $3,844. Since February the campaign has been running on a much smaller budget (~$700-900/month). April is partial through Apr 27 and tracking similar to February/March.

What to do next

1. Fix conversion tracking before anything else. Reported conversions are off by 10-100×. Until the primary conversion is mapped to an actual purchase event (Shopify GA4 → Google Ads, or direct Google tag with a purchase trigger), every PMax optimization decision is being made on bad data. This is the single biggest unlock.
2. Review PMax conversion goals. Specifically check that "Purchase" is the primary goal (and that view-through windows aren't counting page views). Conversion value should equal order revenue.
3. Investigate the 80% conversion value drop. Even with broken counting, $6,840 → $1,398 is a real revenue collapse. Cross-check Shopify orders attributed to Google Ads in March vs April.
4. Consider splitting the account once tracking is fixed. PMax is opaque — for a $700/month budget, a small Search campaign on branded queries would be more transparent and likely more efficient.
5. Check Merchant Center health. PMax leans heavily on Shopping. Disapprovals or feed issues would silently throttle impressions, which matches the 29% impression drop we're seeing.