720768UserReplay Analytics machine learning projectClosedGRD Development of PrototypeInnovate UKWebsites are never perfect. They are complex systems and are prone to technical errors. They are used by human beings, who do not always behave in expected ways. When humans meet software, there will always be a certain amount of trouble. Up to 25% of users struggle when they attempt to make purchases online, and as a result many become frustrated with a website and leave without purchasing. Such struggles are believed to cost the UK economy over £1bn due to abandoned online transactions (source: Experian). Companies are increasingly trying to understand the struggles faced by the customer. Digital Customer Experience Management (Digital CEM) solutions have been developed to address this huge business problem. Digital CEM is based on recording user journeys, storing them for analysis, and enabling journeys of interest to be visually replayed alongside technical diagnostics. The resulting insight allows the online retailer to achieve higher sales by a decrease in the number of abandoned transactions. UserReplay Limited was founded in 2009, and specialises in "session replay technology" designed to allow web developers to record, re-run and analyse a visitor's journey through a website. This service is intended to measure and optimise the digital experience. We are able to find and fix site bugs, resolve disputes, recover abandoned baskets and prevent fraud online. This project seeks to develop a prototype solution which takes the analytical functionality of our existing Digital CEM to a completely new level through the use of data mining and machine learning to automate the process of discovery and diagnosis of customer experience issues, a process which even with the best tools has previously required a highly trained analyst. In doing so, substantial R&D will be necessary to resolve technical uncertainties. The resultant system will provide a step change in capability which will mark it out from existing alternatives.