← Back to projects

E-commerce (outdoor furniture)

The purchase sent by the shop's server: on the same clicks, Google Ads sees 17% more sales.

+17% sales seen
TrackingAds
Chart: purchases counted by the browser versus purchases sent by the shop, August to October 2026
Same months, same campaigns: sales sent by the shop versus sales seen by the browser.

Loungitude sells outdoor furniture online, on PrestaShop. As on any shop, part of the purchases never reached Google Ads: ad blockers, refused cookies, customers closing the payment tab before returning to the site. Bidding was learning from only part of the sales.

Server-side, then server-to-server

I first moved the measurement server-side, through an Addingwell collection server. Then I took the next step: a custom module sends every paid order, from the shop, straight to the collection server, which passes it on to GA4, Google Ads, Meta and TikTok. The customer’s browser is no longer involved.

A comparison on the same clicks

The old browser tag stayed active as a secondary action after the switch. From August to October 2026, on the same campaigns and the same clicks, Google Ads counts:

  • 1,092 purchases through the browser tag;
  • 1,161 through server-side (+6%);
  • 1,280 through server-to-server (+17%).

Checked against the shop’s orders: the server send ties 75% of paid orders to Google Ads, against 64% for the browser, and never exceeds actual sales. The other orders come from other channels.

The result

Since July 2026, bidding relies on the purchase sent by the shop. Depending on the month, Google Ads sees 11% to 28% more purchases, 17% over the three months.

In pictures

Comparison of the three measurement setups: client-side, server-side, server-to-server
The three setups ran side by side on the same account.
Diagram of the server-to-server chain, from the shop to the ad platforms
The chain: the shop sends each order to the collection server, which forwards it to the ad platforms.