This dissertation describes the creation of a new integrated Information Technology (IT) system that assisted in the collection of data about the behaviour of website visitors as well as sales and marketing data for those visitors who turned into customers. A key contribution to knowledge was the creation of a method to predict the outcome of visits to a website from visitors’ browsing behaviour. A new Online Tracking Module (OTM) was created that monitored visitors’ behaviour while they browsed websites. When a visitor converted into a customer, then customer and marketing data as well as sales activity was saved in a new Customer Relationship Management (CRM) system that was implemented in this research. The research focused on service websites. The goal of these websites was to promote products and services online and turn enquiries into offline sales. The challenge faced by these websites was to convince as many visitors as possible to enquire. Most websites relied on Search Engine Optimisation (SEO) and Pay Per Click (PPC) advertising for traffic generation. This research used PPC advertising to generate traffic. An important aspect of PPC advertising was landing page optimisation. The aim of landing page optimisation was to increase the number of visitors to a website who completed a specific action on the website. In the case of the websites investigated in this research the action consisted of completing and sending an enquiry form from the websites. The research looked for meaningful commonalities in the data collected by MS CRM and the OTM and combined this with feedback from the collaborating company’s sales team to create two personas for website visitors who had enquired. Techniques for improving landing pages were identified and these led to changes to landing pages. Some of these changes were targeted at a particular visitor persona. The effect of changes made to a landing page was measured by comparing its conversion rate and bounce rate before and after the changes. Behavioural data collected by the OTM was then analysed using a data mining engine to find models that could predict whether a user would convert based on their browsing behaviour. Models were found that could predict the outcome of a visit to a service website.