Book a call
← All case studies

[ Case 01 · Ecommerce · Google Ads ]

Home and sleep ecommerce: Google Ads spend at 1.8% of real sales

Key numbers

  • 1.8%spend over real sales (May–Aug)
  • US$210Kmanaged in 8 months
  • ~36×channel ROAS validated against real sales
  • 1.8–2.4%acquisition cost as a share of the ticket

The numbers

Google Ads spend as % of real sales, May–Aug 2026

Four-month average: 1.81%. Lower is better.

Channel ROAS by source of truth

The case works with the last one.

The account

A Peruvian ecommerce selling mattresses, beds and bedroom furniture, online and in physical stores. Average ticket around US$350. Spend managed: US$210,224 over eight months, January to August 2026.

The problem

Efficiency had been eroding for six months. Platform ROAS fell from 85× to 28× and cost per acquisition rose from US$4 to US$13. Two causes: competitive pressure on the brand term and Performance Max cannibalizing Search.

What I did

  1. April: restructured Performance Max into themed campaigns by product line and added high-intent Search. Adjusted the brand campaign’s target ROAS from 35 to 25.
  2. June: separated Search and Performance Max with brand negatives to stop the cannibalization.

The result

From May to August, Google Ads spend equaled 1.8% of real closed sales. Channel ROAS came to about 36×, validated against the payment gateway. July was the best month in the account’s history.

How to read this case

  • Spend over sales is Google Ads spend divided by real closed sales in the client’s payment gateway. It’s not platform attribution; it’s what the client actually billed. GA4 attributes 65% of those sales to the channels I managed.
  • Which ROAS. This is ecommerce with measurable revenue, so ROAS is valid. But I discard the 60× Google reports: it over-attributes about 33% of purchases compared with the client’s analytics. I use GA4’s ~40×, adjusted to ~36× against real sales.
  • Gross vs. net. GA4 records the purchase at checkout; the gateway records the settled sale. The 2–15% gap each month is cancellations and returns. The case reports net.
  • Brand included. Brand defense is 31% of spend and stays in. Excluding it would push apparent ROAS to about 50×, and that would be a convenience cut.

Breakdown by channel, Jan–Aug 2026 (platform data)

See the full table
ChannelSpendPurchases (platform)Platform ROASShare of spend
Search (brand + non-brand)US$82,42315,84574×39%
Performance MaxUS$69,61810,34654×33%
ShoppingUS$50,5777,05152×24%
Display (remarketing)US$7,606915×4%
TotalUS$210,22433,33360×100%

Spend over real sales, May–Aug 2026 (payment gateway)

See the full table
MonthGoogle Ads spendReal salesSpend / sales
MayUS$26,885US$1,346,9202.00%
JuneUS$29,694US$1,205,1352.46%
JulyUS$40,583US$3,092,3981.31%
AugustUS$28,514US$1,304,8862.19%
TotalUS$125,676US$6,949,3401.81%

Source and scope

Spend from Google Ads (USD), one account, Peru market. Revenue from GA4 and the client’s payment gateway. Real sales are available for May–August only; January–April is reported at platform and GA4 level. Meta ran in parallel; Google drove 96% of paid-channel revenue in GA4. Exchange rate S/3.60 per USD. Client and keywords anonymized.

Next case →−45% cost per quality lead, same budget

Let’s find out what your numbers actually say.