Enhancing the customer experience
Recommended product purchases have doubled across the online merchants' superstores.
Who are you?
CMOStores.com is a group of six specialist e-commerce stores, dedicated to providing our customer base – a mixture of DIYers, traditional trade and non-trade businesses – with the building supplies and construction materials they need.
We’re an award-winning and forward-thinking e-commerce business that’s committed to embracing automation and innovation in the builders’ merchants sector. This year, we turned to Decision Intelligence and AI to revolutionize our customers’ experience, optimizing our six websites to enhance the customer experience significantly.
What was your challenge?
As certain building or DIY projects require a particular combination of products, we needed a solution that allowed our customers to get to work quicker and with less hassle. We wanted to make the process of searching for the right products much simpler, while also reminding users of associated products they might have forgotten to add to their basket.
In addition, with industries such as construction gripped by uncertainty during the early stages of the pandemic, we had to provide quick and easy access to the products our customers needed to support projects and keep Britain building during unprecedented times.
We want to make sure that our customers are connected with all the products they need to complete their projects, and have the best experience when shopping with us.
Marketing Director, CMOStores.com
What did Peak do?
Peak implemented an AI-powered solution to help us gain a clearer understanding of our customers’ behaviour. CODI, Peak’s Decision Intelligence system, analyzes daily transactional data from our data warehouse, and identifies the most relevant product recommendations for customers based on their past purchasing habits.
This data is ingested into CODI on a daily basis, following which the recommender model is trained and recommendations generated. These insights are automatically uploaded back into our website via an API hosted on CODI and are then served to users as dynamic recommendations. This AI-powered view enables us to provide highly relevant, personalised product recommendations across our online superstores.
In the six months since we partnered with Peak to create hyper-personalized product recommendations at the point of sale across our webstores, CODI has driven twice as many transactions as our original non-AI approach.
What’s the upshot?
The implementation of AI-driven recommendations has allowed us to provide our customers with a significantly improved product experience, making it easier and more straightforward for shoppers to find and purchase the items they need. In turn, customers are able to increase the efficiency and speed of whatever building or DIY project they’re working on.
Thanks to Peak, we achieved a rapid time-to-value amidst COVID-19 and, in fact, saw our market share increase this summer as more homeowners shifted to online purchasing, and demand for construction products rose thanks to a nationwide trend for DIY projects during lockdown.
Over a fifth of all transactions on the sites have involved an interaction with CODI, proving that customers are using and benefiting from the new technology. Our overall website session values increased by 11.2% compared to our previous non-AI approach, with our total revenue rising by 5%. Peak’s recommendations were clicked on by 60% more customers, and had a 25% higher click-to-transaction rate.
Peak makes sure we can offer the right product to the right person at the right time – keeping customers at the heart of our business while using exciting new technology to deliver real benefits to them.
Marketing Director, CMOStores.com
Under the hood: How Amazon Web Services (AWS) technology underpins this solution
As an AWS Machine Learning and Retail Competency Partner, Peak’s solutions are powered by a range of AWS services. In this instance, data is ingested from CMOStores.com into Amazon Redshift, which is then processed and the recommendation model trained on Amazon EKS. This is then served back to CMOStores.com via an Amazon SageMaker endpoint. All development work in this use case was carried out on AWS Elastic Beanstalk.