What the work looks like - and the thinking behind it
Growth is a marathon, not a sprint - a bad week, even a bad month, is fine. Do a data review and fine-tune the media; don’t rebuild everything and drag the systems through a new learning phase.
In Shopping, remember the product feed is often the first thing a shopper actually sees, so give it the respect it deserves: update it as often as customer needs change.
CRO and a smooth buying experience matter as much as the media - everything after the click counts too, all the way to a customer who comes back.
A discount can sell, but there are better, longer-term ways: understand what matters to your personas and give them that value, because “15% off” isn’t a value - it’s just what everyone else does.
Google Ads · E-commerce+38% ROAS*whole account, January to January
GA4Google AdsMerchant CenterShopping
More of the auction, more sales, less budget
The product feed was rebuilt so every item matched the searches it should actually win - generic data had been pulling us into auctions for products we don’t sell, burning budget on queries that could never convert. Shopping position came from relevance, not bids.
The two-month drop is a deliberate pause, not lost ground.
Multi-channel · E-commerceMore than 100% ROAS*blended, January to January
Meta AdsGoogle AdsMerchant CenterShoppingGA4ClarityShopify
Less budget, more sales - the same account a year later
Media budget was cut on purpose year over year, with funnel-wide optimization protecting performance on both platforms regardless. A secondary but real factor: GA4 and Microsoft Clarity showed homepage offers were barely seen, and fixing that gave both platforms cleaner signal too - the sales that followed fed richer funnel signal back into both algorithms, each side compounding the other.
Sharper targeting lifted Meta’s CTR, cutting CPC and CPA through more qualified traffic. Google shifted spend from broad country coverage to the countries that were actually profitable.
Blended · January vs January
Meta · January vs January
Google · January vs January
Google Ads · E-commerce+72% sales
Trading ROAS for revenue on purpose
Budget more than doubled for peak season, and revenue climbed right along with it - the extra volume did trade some ROAS for higher CPCs, but conversions and revenue still scaled to match the spend.
Meta Ads · E-commerce-21% CPC
Cheaper clicks as the audiences got sharper
Audiences rebuilt around purchase intent, creative refreshed by segment, campaigns split by product line instead of one catch-all set.
cost per clickclick-through rate
Google Ads · E-commerce-6% CPA
Small budget increase, broad gains across the funnel
Search and Shopping paced against margin rather than a blanket target, feed and bids tuned as the season moved.
Q2 vs year prior
Google Ads · Full account+26% ROASwhole account, this half vs last
Google AdsGA4
Six months in, every lever moved the same way
Search structure was rebuilt around intent instead of one catch-all set, negative keywords were cleaned monthly, and ad copy testing followed what GA4 showed was actually converting on-site. CTR climbed first, and Quality Score did the rest - a stronger Ad Rank pulled CPC down even as spend kept rising into the auction.
Conversion rate followed once traffic quality caught up, and only then did the bid strategy move onto a smarter automated target, by which point conversion volume gave it enough signal to work with. None of these numbers moved in isolation - each fed the next.
CPC · conversion rate, before vs after
Account · this half vs last
+7%cost
+16%clicks
+42%conversions
+26%ROAS
1CTR climbs first
2Quality Score, then Ad Rank
3CPC pulled down
4Conversion rate follows
5Bid strategy goes automated
Multi-channel · Customer acquisition>100% new customersQ4 vs the Q4 before it
Meta AdsGoogle AdsGA4Shopify
A Q4 built to grow the customer file, not just the cart
Budget across the paid mix scaled up for the season and was weighted toward prospecting instead of leaning on remarketing - Meta shifted onto lookalikes built from high-value past purchasers, Google moved spend into Performance Max’s acquisition signal, and creative was tested cold rather than reused from retargeting sets. Shopify and GA4 confirmed the lift was genuinely new buyers, not the same customers shopping a seasonal sale early.
Cost per new customer fell even as both spend and the count of new customers surged - that combination is the tell. More budget alone buys volume at a flat or rising cost; this was the acquisition funnel itself converting more efficiently.
New customers
Cost per new customer
Multi-channel · E-commerce-6% CPAblended, this half vs previous
Meta AdsGoogle AdsShopifyGA4
Flat budget, a materially bigger half year
Creative testing across both platforms surfaced which angles actually drove purchases rather than just clicks, and the Shopping feed was rebuilt for tighter title-to-query matching so budget stopped landing on searches that were never going to convert. Remarketing was also split by drop-off point, instead of showing the same generic offer to a browser and a cart-abandoner alike.
Cost held essentially flat across the half - this wasn’t a budget increase buying the extra revenue, it was the same spend working harder once the mechanics changed.
Revenue · ROAS, before vs after
Charts describe the change in an account, not the account itself. Figures are period over period from each account’s own reporting. *Percentages are rounded down beyond ordinary rounding, specifically to protect confidentiality - actual results are stronger than shown. Client names, brands and markets are withheld deliberately.