05 · Automotive brand (confidential)
Scaled Google Search + Shopping past $300K/month at 4×+ ROAS.
2022 – present · Growth Marketing Lead
The bet
Most brands run Google like it's 2019.
Launch a Smart Shopping campaign. Set a budget. Walk away.
Maybe layer in brand-defense Search if they're feeling ambitious.
I bet we could run Google like Meta. Test angles. Segment audiences. Compound winners.
The unlock was the feed.
Most brands ship a flat Shopping feed: raw product titles, one image, no segmentation.
I rebuilt the feed with custom_label_0 through custom_label_4.
Each SKU tagged by tier, margin band, seasonality, top-3 competitor, and traffic-source intent.
Then I split Performance Max into four campaigns by tier (custom_label_0).
Scale high-margin advanced-tier campaigns aggressively. Don't burn budget on low-margin entry models.
Search got the same treatment.
Branded campaigns ran defensive. Non-brand campaigns segmented by buyer intent: make/model queries in one bucket, problem/solution queries in another.
But the real ROAS unlock wasn't in the account.
It was on the landing page.
Most brands send every Search click to the same PDP. I built a dedicated landing page for each hero product.
Match the keyword intent to the page. Specific headline, specific benefits, specific proof. No friction, no menu noise.
Hero-product LPs lifted ROAS by 250% over the default PDP.
For the undecided buyers (the ones who searched broad category terms), I built a quiz funnel.
Three questions, recommend the tier, send them to the matched LP.
Then I fixed GA4.
The default GA4 attribution model was undercrediting Google by 30–40% versus the proxy events Amazon could actually verify.
Reconfigured conversion paths, plugged offline conversions back in, swapped the attribution model to data-driven. Suddenly Google's ROAS reads aligned with what we saw on the P&L.
Six months in, monthly Google spend was past $300K at 4×+ ROAS. Held through Q4.
Concept
The Google account mix
Every campaign type serves a different intent. Mixing them right is what unlocks scale without cannibalizing brand.
- 01
Shopping
Feed-first. Custom labels segment by tier, margin, season.
- 02
Performance Max
Split by tier label. Asset groups stay siloed by margin band.
- 03
Non-brand Search
Segmented by buyer intent: make/model vs problem/solution.
- 04
Brand defense
Run hot. PMax tries to eat brand if you don't fence it off.
- 05
Demand Gen
YouTube + Discovery. Top-funnel layer feeding the rest.
Concept
Three pieces that unlocked the 250% ROAS jump
The ad gets the click. These get the conversion. Most brands skip them entirely.
- 01
Hero-product landing pages
One LP per hero SKU. Headline matches search intent. Single CTA. +250% ROAS over default PDP.
- 02
Quiz funnel
Three questions for the undecided buyer. Recommends the right tier. Routes to the matched LP.
- 03
GA4 attribution rebuild
Data-driven model. Offline conversions reconnected. Reads finally aligned with the P&L.
Concept
Feed-first scaling
The feed is the strategy. The labels are the levers.
- 01
Feed audit
Identify missing attributes, gap fields
- 02
Custom labels
Tier · margin · season · competitor · intent
- 03
Segmented PMax
4 campaigns split by custom_label_0
- 04
Brand fenced off
Brand kw excluded from PMax, defensive Search
- 05
Scale tier by tier
High-margin tiers get the budget
What broke
Performance Max wanted to eat my brand traffic.
Free conversions, free attribution, easy ROAS. Google's algorithm loves the lazy win.
I had to exclude brand keywords from PMax AND bump bids on the brand Search campaign to keep PMax from poaching it.
You have to fence off the easy conversions if you want a real read on non-brand performance.
The result
The feed is the strategy. The landing page is the closer. The attribution is the proof. Get all three right and Google scales without theatrics.
Next
Want me to run a similar play for you?
Open to senior growth roles and select advisory work.