I Spent £2M on Google Ads for Ecommerce. Here's What Actually Works

Over £2 million spent on Google Ads. 100+ ecommerce stores managed. Average ROAS: 4.2x. Here's what works, what doesn't, and what most agencies won't tell you.

Let's Start With What Doesn't Work

Before I tell you what to do, here's what to stop doing immediately:

1. Broad Match Keywords Without Negative Lists

I took over an account spending £8,000/month on broad match keywords with zero negative keywords. They were showing ads for "free watches" and "how to make jewelry at home." Burned £2,400 in one month on completely irrelevant clicks.

Lesson: Broad match can work, but only with aggressive negative keyword management.

2. Using Google's "Maximize Conversions" Without Constraints

Google will happily spend your entire budget to get you conversions—even if those conversions are unprofitable. I've seen accounts with 3.2x ROAS celebrating "increased conversions" while losing money on every sale.

Lesson: Always set Target ROAS or Target CPA. Never give Google a blank check.

3. Display Campaigns for Cold Traffic

Display ads can work for retargeting. For cold traffic? I've tested it 47 times across different industries. Average ROAS: 1.1x. Not worth it for most ecommerce stores.

Exception: If you're selling impulse-buy products under £20, display can work. Otherwise, skip it.

4. Single Product Ads for Low AOV Products

Running individual ads for £15 products is a waste. With typical CPCs of £0.80-£2.00, you need a 10%+ conversion rate just to break even. That's unrealistic.

Lesson: Focus on product bundles, upsells, or higher-margin items.

Now, Here's What Actually Works

Strategy #1: Shopping Campaigns First, Search Second

Why: Shopping ads show your product, price, and image. Self-qualifying traffic. Lower CPCs. Higher intent.

Data from my accounts:

• Shopping campaigns: Average ROAS 5.8x
• Search campaigns: Average ROAS 3.2x

Budget allocation rule: 70% Shopping, 30% Search (adjust based on performance, but start here)

How to Set Up Shopping Campaigns Right:

1. Product feed optimization is everything
   • Use all available characters in titles (150 characters)
   • Front-load with keywords: "Leather Laptop Bag Men 15 Inch" not "Premium Executive Bag"
   • Include brand, color, size, material in every title
   • Add custom labels (margin, bestseller, seasonal, clearance)
2. Campaign structure matters
   • High Priority Campaign: Negative keywords to exclude broad terms, bid £0.01 (just to capture data)
   • Medium Priority Campaign: Some negative keywords, moderate bids
   • Low Priority Campaign: No negative keywords, highest bids

   This structure gives you granular control over search terms and bids.

3. Use Performance Max (but carefully)
   • Performance Max works, but you lose control
   • Start with 20-30% of Shopping budget in PMax
   • Monitor search term reports religiously
   • Add negatives at the account level to control PMax

Strategy #2: Brand Defense Campaigns (Non-Negotiable)

The situation: Your competitors are bidding on your brand name. If you don't, you're giving them free traffic.

Real example: Client refused to bid on their own brand name. Competitors were getting 30% of brand search traffic. We launched a brand campaign, spent £380/month, protected £14,000/month in brand revenue.

How to do it:

• Create a separate campaign for your brand terms
• Bid high enough to stay in position 1
• Use site link extensions to control the message
• Monitor impression share—you want 95%+ for brand terms

Strategy #3: SKU-Level Bid Optimization

Not all products are created equal. Some have 60% margins, some have 15%. Some convert at 8%, some at 1.2%.

What I do:

• Export all product performance data
• Calculate profit per product (not just revenue or ROAS)
• Create custom labels: "High Margin," "Low Margin," "Bestseller," "Slow Mover"
• Set different bids for each label

Example:

• High margin products (50%+ margin): Bid £2.50
• Medium margin (30-50%): Bid £1.20
• Low margin (under 30%): Bid £0.40 or pause

Result: Overall ROAS increased from 3.8x to 6.2x just by bidding smarter.

Strategy #4: Negative Keywords Are Your Best Friend

I have a negative keyword list with 1,200+ terms built over 10 years. Every account I manage gets this list immediately.

Common negative keywords for ecommerce:

• "free"
• "cheap"
• "how to make"
• "DIY"
• "jobs"
• "wholesale"
• "repair"
• "used"
• "second hand"

Process:

• Check search term reports every 3 days
• Add irrelevant terms as negatives immediately
• Use negative keyword lists at the account level (applies to all campaigns)

Impact: Average CPC reduction of 18-25% after proper negative keyword management.

Strategy #5: Retargeting with RLSA (Remarketing Lists for Search Ads)

Someone visited your site, didn't buy, then searches for your product category again. That's a hot lead.

How RLSA works:

• Create audience lists (site visitors, cart abandoners, past purchasers)
• Layer these audiences onto your search campaigns
• Bid higher for people who've already visited your site

My RLSA strategy:

• Standard bid for cold traffic: £1.00
• Bid for site visitors (last 30 days): £2.20 (+120%)
• Bid for cart abandoners: £3.50 (+250%)

Results:

• RLSA campaigns average 9.4x ROAS
• Conversion rates 3-5x higher than cold traffic

Strategy #6: Dynamic Search Ads for Long-Tail Keywords

If you have 500+ products, you can't possibly create ads for every variation. Dynamic Search Ads solve this.

How they work:

• Google crawls your website
• Automatically generates ads based on page content
• Shows ads for relevant searches you haven't targeted

Best practices:

• Start with 10-15% of total budget
• Exclude pages you don't want to advertise (about page, blog, policies)
• Monitor search terms closely—add good terms to regular campaigns
• Use negative keywords aggressively

Results: DSAs typically capture an additional 15-20% of revenue beyond standard campaigns.

Strategy #7: Seasonal Bid Adjustments

Your business isn't the same every month. Neither should your bids be.

What I track:

• Conversion rates by month
• Average order value by month
• Competition levels by month

Example (fashion retailer):

• January: -30% bids (low conversion, post-Christmas slump)
• March-April: +20% bids (spring collection launch)
• July: -20% bids (summer slowdown)
• November-December: +50% bids (Black Friday, Christmas)

Impact: 23% improvement in yearly ROAS by matching bids to seasonal demand.

The Technical Setup Most People Get Wrong

Conversion Tracking (It's Broken on 40% of Stores)

I audit 5-10 Google Ads accounts every month. 40% have broken or incomplete conversion tracking.

What to check:

• Is Google Ads tracking all purchases? (Check in Google Analytics to compare)
• Are you tracking order value, not just order count?
• Is your tracking firing on the thank you page (not the checkout page)?
• Are you excluding internal traffic?
• Are you using enhanced conversions?

How to test: Make a test purchase and see if it shows up in Google Ads within 24 hours with the correct order value.

Enhanced Conversions (Must Have for Privacy Changes)

With iOS privacy updates and cookie restrictions, standard conversion tracking is less reliable. Enhanced conversions help fill the gaps.

What it does: Sends hashed customer data (email, phone) to Google to match conversions even when cookies are blocked.

How to set it up:

• Use Google Tag Manager
• Implement enhanced conversions tag
• Test with Tag Assistant

Impact: 15-20% more accurate conversion data (huge for optimization).

Budget Allocation Framework

Here's exactly how I allocate budget for a new ecommerce account:

Month 1-2 (Learning Phase):

• 60% Shopping campaigns
• 20% Brand search
• 10% Dynamic search ads
• 10% Testing (RLSA, Performance Max, etc.)

Month 3+ (Optimized Phase):

• 50% Shopping campaigns
• 15% Performance Max
• 15% Brand search
• 10% RLSA
• 10% Dynamic search ads

Budget minimums:

• Don't run Google Ads with less than £500/month total budget
• Each campaign needs at least £10/day to gather meaningful data
• Shopping campaigns need £20/day minimum

When to Pause or Kill Campaigns

Not everything works. Here's when to pull the plug:

• After 2 weeks, no conversions: Pause and re-evaluate
• ROAS below your break-even for 30 days: Pause or restructure
• CPC increasing without conversion rate improvement: Investigate and likely pause
• Impression share dropping below 50%: You're not competitive, increase bids or pause

Real Results From Recent Clients

Client 1: Home Decor Store

• Starting point: £2,800/month ad spend, 2.1x ROAS
• Changes: Shopping campaign restructure, negative keywords, SKU-level bidding, RLSA
• Results after 90 days: £4,200/month ad spend, 6.8x ROAS
• Revenue increase: £8,500/month additional profit

Client 2: Fashion Accessories

• Starting point: £6,500/month ad spend, 3.4x ROAS (barely profitable)
• Changes: Performance Max + Shopping combo, enhanced conversions, seasonal bid adjustments
• Results after 90 days: £8,200/month ad spend, 5.9x ROAS
• Revenue increase: £22,000/month additional profit

The Truth About Agencies

Most Google Ads agencies are incentivized to spend more, not to make you more profitable. They get paid a % of ad spend—so they want bigger budgets.

What I do differently:

• I charge based on results, not ad spend
• If I can get you the same revenue with less ad spend, I do that
• I focus on profit, not vanity metrics like impressions or clicks

Want Me to Audit Your Account?

I offer free Google Ads audits where I'll:

• Review your current campaign structure
• Identify wasted spend
• Show you untapped opportunities
• Estimate the revenue impact of fixing issues