Maximizing Revenue for a Sports Retailer with a +100k Product Catalog

Driving Budget Efficiencies & Automating Feed Optimization

Our client, a worldwide sports retailer with 38 stores in Portugal, wanted to use Google Merchant
Center feed to increase revenue and sales volume on their e-commerce website. The brand sells over
100,000 products for over 100 sports, managing multiple promotional campaigns and discounts
dedicated to different sports and seasons throughout the year.

OUR BOLD AMBITION

With a large product inventory and ambitious KPIs to hit, our client needed a solution to help them stay
ahead of the competition. They had to be ready to adapt and embrace new technologies to maximize
the ROI of both traditional shopping campaigns and new Google Performance Max (Pmax) campaigns.
The brand was facing e-commerce obstacles, such as inconsistent inventory levels during the pandemic,
and macroeconomic challenges, including inflation and its impact on consumer spending – and it was
more important than ever for us to improve efficiencies and performance. Our goal was to enhance our
client’s e-commerce success by expanding the products’ points of sale to multiple channels and
automating processes using machine learning.

OUR SOLUTION

Our team structured the project into three main phases:
‍
1. Analyzing campaigns’ performance
2. Creating a supplemental feed to exclude a set of products
3. Leveraging Google Cloud to manage data
‍
We started by analyzing product performance data on Pmax campaigns to identify which products were
unprofitable and should be excluded from the feed to drive efficiencies and channel the current budget
into more profitable products.
To identify which products to exclude, we set criteria based on high impressions, low CTR and low ROAS,
which were in line with the client’s KPIs. We then created a supplemental feed to provide additional
attributes that were missing or not included in the primary feed. This feed contained information about
the list of products to exclude and will contribute to increase the effectiveness of the brand’s Pmax
campaigns.

We used Google Cloud functions to power this process and BigQuery to manage not only performance
data but also product-level insights. These insights will enable our client to continue improving activities
and reintroduce products into Pmax campaigns as demand and conversion rates increase.

OUR RESULTS

We developed the following matrix model to analyze feed optimization performance. The matrix helped
us identify if the strategy applied provided an opportunity to deliver greater results. Each plot point
represents the combination between ROAS and revenue that we achieved during the test.

By leveraging data efficiently with machine learning, we boosted overall performance on Pmax campaigns. We drove budget efficiencies by reducing costs by 15%, while simultaneously producing a 9% increase in revenue and a 29% in ROAS.