If you’re running paid media for an e-commerce brand and still lumping all your products into one Performance Max campaign, you’re leaving serious money on the table.

I recently helped an e-commerce client increase their return on ad spend (ROAS) by 57% in just over two months – without increasing total spend. The secret? Smart segmentation, precise targeting, and giving Google’s algorithm the right signals to work with.

Let me walk you through the exact strategy we used. Each element had a purpose. Each campaign was built to serve a specific intent. And everything was aligned to drive profitable growth, not just traffic.

1. Performance Max – Focused on Top Performers Only

We started by splitting out best-selling products into their own PMax campaign.

• The product feed was filtered to include only the highest converting SKUs.
• Budgets were prioritised for these proven winners.
• We stripped out everything but Shopping ads, keeping things tight to maximise efficiency.

By giving the algorithm clean, high-intent data and a narrowed focus, it delivered consistent, scalable returns. Too often, advertisers let PMax run wild across too many SKUs. Focus changes the game.

2. Search Campaign Built on Amazon-Like Queries

We mirrored the kind of high-intent, purchase-ready terms that users type into Amazon.

• Think: “[product type] for [specific use case]” or “[brand] + [feature]”.
• We pulled insights from the client’s Amazon listings to find exactly what was driving conversions there, and brought those terms over to Google.

It’s a move that helped us capture in-market buyers while competitors were still targeting broad, generic terms.

3. Branded Search – Exact and Phrase Match Only

We locked down branded traffic using only exact and phrase match keywords.

• This let us control messaging, manage ad extensions, and avoid inflated CPCs from broad match leakage.
• It also made sure we weren’t bleeding conversions through affiliate sites or competitors bidding on our name.

If you’re not protecting your brand terms properly, you’re probably leaking performance without realising.

4. Performance Max – Low Performers in Their Own Box

Next, we took the slower-moving SKUs and housed them in a separate PMax campaign.

• These had tightly controlled budgets.
• We used them primarily for data collection and long-tail discovery.
• Again, we limited it to Shopping ads only to keep the focus on purchase intent.

The aim here wasn’t to scale – it was to gather signal and find potential hidden gems without letting them drag down our top campaign.

5. Non-Brand Search – Segmented by Funnel Stage

We built our non-branded search campaigns with intent in mind.

• Top-of-funnel ad groups had educational messaging, softer CTAs, and blog or category page landers.
• Bottom-of-funnel groups drove users straight to product pages, with hard-hitting copy and urgency.

Each intent layer had its own ad copy, its own bids, and its own landing pages. That structure made sure we weren’t showing the wrong message to the wrong person.

6. Dynamic Search Ads (DSA) – Catch-All for the Unknown

DSAs were our safety net.

• We opened up to all pages across the site.
• It helped us pick up unexpected search terms and uncover demand we weren’t specifically targeting.
• Most importantly, it fed the keyword mining process for future campaigns.

Don’t underestimate DSAs – when used with a clean site structure, they can do a lot of heavy lifting behind the scenes.

7. Geo-Targeted Performance Max

Rather than going broad with budget, we zeroed in on one high-performing region.

• We used geo-fencing to isolate a specific area where ROAS was already strong.
• That localisation gave Google clearer data signals and more relevant user behaviour to optimise around.

Especially for DTC brands with varied regional demand, this kind of geographic focus can unlock far more efficient scaling.

8. Display Retargeting – With a Strategic Twist

Retargeting wasn’t just about reminders.

• We used Display campaigns specifically to re-engage warm users with high intent.
• Every creative included urgency, reviews, and limited-time offers – not just a picture of the product.

It wasn’t about running a passive loop of ads. It was about re-selling the value with purpose.

9. Branded Shopping Campaign

Finally, we pulled branded Shopping queries into their own campaign.

• These targeted exact brand and product name combinations – the kind of searches that almost always convert.
• Cost-per-clicks were low, and ROAS was consistently high.

Most brands overlook this. But branded Shopping is often the cheapest, easiest win you can grab.

Why It Worked

We didn’t reinvent the wheel. But we did do the fundamentals really well:

• Campaigns were structured around buyer intent.
• Budgets were allocated based on proven ROI, not guesswork.
• Every product and query had a home – nothing was left to chance.

By feeding Google clean data, clear signals, and tightly defined goals, the algorithm did exactly what it’s supposed to do: scale what works.

The Bottom Line

This strategy lifted ROAS by 57% in just over nine weeks. No hacks. No gimmicks. Just disciplined execution, proper segmentation, and a deep understanding of how Google’s machine learning works.

If your Google Ads strategy still looks like one PMax campaign and a couple of broad search terms, it’s time to raise your game. You don’t have to spend more to scale smarter – you just need to be more intentional with how you set up your campaigns.

Thinking of overhauling your Google Ads structure? Let’s have a conversation.