MERCK // CHATBOT ASSISTANT

BUSINESS PROBLEM

Multinational pharmaceutical company, Merck, sought to increase user engagement of their oncology resource website. By enriching the site experience, they aimed to build more credibility around the Merck oncology product portfolio. As the lead strategist assigned to the project, I conducted a thorough audit of the site content, user journey, engagement analytics, and user testing results in order to provide optimization recommendations. The majority of site visitors were over the age of 45, and user testing results revealed challenges in navigating current site content. The site needed a better way to assist users and create a more personalized experience.

SOLUTION & APPROACH

In addition to redesign and content hierarchy optimizations, I recommended developing a custom chatbot tool to help boost engagement and to provide personalized content recommendations for users. Several third-party studies show that chatbots are a more reliable and accurate alternative to medical condition-specific organic searches. A chatbot could help address the needs of two target audience segments on the site: patients seeking ways to improve their health and manage diseases like diabetes, cancer, and cardiovascular disease, and caregivers looking to deliver better care to their patients and enhance their own wellbeing.

Leaning on the analytics and testing data, I organized the available content into conversational flows, often employing buttons and emojis to quickly guide user inquiries. The recipe content, in particular, lacked connection to specific health conditions so I created a system of tags and filters that would allow users to easily explore recipes based on their dietary requirements, which would further enable better disease management.

I worked closely with a UX designer to create an intuitive, responsive interface. I also collaborated with a creative technologist and development team to capture backend analytics and incorporate natural language processing into the chatbot functionality as a shortcut option for users. After rounds of wireframing and prototyping with our cross-functional team, I created a user testing plan and script to uncover any potential roadblocks before launching the chatbot. Upon analyzing the new testing results, I added a final question to the chat experience to collect voluntary user sentiment and feedback. This feature would help our team further improve the chatbot in future iterations. We successfully launched Merck’s first CMS-driven AssistME chatbot, the client was over the moon, and the website experienced an overall uptick in engagement and traffic directed to Merck product pages.