AI e-commerce analysis is becoming essential for retailers who want to understand what truly drives basket value, checkout completion, and customer purchasing behaviour. In this case study, we show how Trimontium transformed complex shopping-basket data into clear, actionable insights that revealed the factors influencing conversion and revenue growth.
Why AI Analysis of Website Sessions Was Critical for our Client
Impact: Optimised cross-selling increased average basket size by 22% and average revenue by 36%
Case Study: AI e-commerce analysis of shopping basket behaviour
Client sector: E-commerce Client
Our Goal: To turn unusable basket data into actionable intelligence that explains what boosts basket value, what drives checkout completion, and how to increase revenue.
A growing e-commerce business approached Trimontium with a challenge. Their online store generated large volumes of shopping basket data — recording which items users added, which baskets were completed, and which were abandoned. While they knew this data was valuable, they had no clear way to extract insight from it. They wanted to understand what drove basket completion, what made some baskets more valuable than others, and how to design experiences that encouraged users to buy more. However, with huge amounts of complex, unstructured data and no analytical framework, they were unable to identify patterns, test ideas, or act on their findings. The data existed but it wasn’t usable for decision-making.
Trimontium’s approach
We knew that the answers lay hidden in the client’s existing data. However, uncovering those insights required advanced statistical and machine learning techniques far beyond what manual analysis or standard reporting tools could deliver. Our goal was to bring all of the client’s e-commerce data together within the Trimetrics platform, then use Trimontium’s AI-driven analytical engine to model and explain what truly influenced basket value and conversion. The result would be a data-driven understanding of what worked, why it worked, and how to act on it.
What we built
Trimontium first built automated daily pipelines to collect and unify all of the client’s transactional and behavioural data within the Trimetrics platform. Once the data was consolidated, we applied a suite of statistical and machine learning models to explore two key questions:
What features are associated with high-value baskets?
We analysed both existing and derived features, including session duration, traffic source (organic or paid), number of interactions, and product mix. Our models assessed the relative importance of each factor in predicting basket value and completion likelihood.
How does the client’s data compare to wider patterns across sectors?
By benchmarking the results against Trimontium’s aggregated cross-industry models, we could determine which behaviours were specific to this client and which reflected broader market trends, giving valuable context for interpreting results.
In parallel, we designed the foundation for a personalisation engine, capable of recommending additional products based on basket composition. For example, if a user added product A to their basket, the system could suggest complementary items B or C in real time, a natural next step made possible by the insights from our analysis.
Outcomes of Trimetrics’ AI e-commerce analysis
The client gained deep, actionable understanding of what drove both basket value and checkout completion. For the first time, they could see which factors such as user behaviour, product combinations, or traffic source had the greatest impact on sales outcomes. Armed with these insights, the client began to refine their on-site experience, pricing, and cross-sell strategy. Trimontium’s AI models also opened the door to intelligent personalisation, enabling the business to tailor recommendations dynamically and increase conversion rates. By turning raw basket data into a live AI system, Trimontium empowered the client to make data-driven decisions that boosted revenue, improved customer experience, and set the foundation for continuous optimisation.
Trimetrics and the AI and data science solutions were designed, implemented and optimised by Trimontium’s Data Science Team, led by our Chief Data Scientist, Dr Peter Appleby.
