Fueling AI Excellence

Sean MacCarthy, VP of Analytics at Camping World/Good Sam, began his career as a philosophy professor. Later, he transitioned to the corporate sector, marking the start of a data-driven journey where he embraced solving business challenges through analytics and data science. In the fast-paced business environment, Sean quickly adapted, mastering the ability to address complex problems within tight timeframes.

His career has been dedicated to leveraging data science across critical functions such as operations, marketing, pricing and supply chain management. At CW/Good Sam, Sean specializes in decoding the complexities of customer behavior across the whole Recreational Vehicle (RV) lifestyle. His understanding of human unpredictability allows him to blend overarching trends with personalized, data-driven solutions. By employing advanced analytics and cutting-edge digital tools, Sean ensures the delivery of precise, customer-centric innovations, helping CW/Good Sam maintain its competitive edge.

Given your extensive experience, what are the key challenges in today’s market?

One of the most persistent challenges across companies is data quality—having clean, well-governed data that is easily accessible to analysts, operators and marketers. While some organizations excel at this, many struggle due to fragmented tech stacks that combine outdated legacy systems with newer technologies that fail to integrate seamlessly.

No matter how advanced or hyped tools like large language models (LLMs), generative AI or machine learning are, their effectiveness is limited without high-quality data. Poor data governance can lead to biased models, inaccurate results and misaligned customer targeting. To fully leverage AI capabilities, organizations must prioritize robust data stewardship and governance.

Industries like finance and healthcare have made strides in governance for security and legal requirements. Retail, where governance practices are still maturing, must now focus on supporting data scientists and AI teams. Treating data as a product, not just a regulatory obligation, will be vital over the next 5–10 years to unlock the true potential of AI and advanced analytics.

What trends or technologies do you consider key to addressing these challenges?

Addressing data challenges starts with strong processes. Implementing robust, systemic processes ensures clean data and sustainable practices. This includes building healthy system architectures when updating or integrating new technologies. Engaging with data teams and leadership helps identify pain points, such as excessive time spent on data engineering instead of leveraging data for strategic solutions. Treating data as a product—complete with dedicated product ownership—ensures long-term advocacy for data cleanliness and compliance.

 AI will live up to its potential in areas like skew rationalization and planogram recommendations, empowering merchants and streamlining analysts’ work. 

For example, call centres can benefit from process updates to improve compliance, like ensuring dropdowns are filled correctly. Automating parts of these tasks, such as using AI technologies like voice-to-text, can streamline operations while enabling associates to focus more on customers.

Voice-to-text technology has immense potential, offering insights into customer sentiment, coaching opportunities and identifying successful communication patterns. By pairing it with generative AI and LLMs, companies can analyze interactions efficiently, whether in customer service, sales or even B2B scenarios like software support.

Can you share a recent project where you’ve implemented these trends or technologies to drive success?

We’ve been exploring voice-to-text technology across various areas of the company, particularly for customer churn modelling, which has proven highly effective. On the analytics side, we’ve begun leveraging LLMs to enhance problem-solving efficiency. For instance, when tasked with analyzing complex datasets, such as hundreds of stores and dozens of product lines—an analyst can now use an LLM to automate parts of the analysis. With prompt refinement, even less experienced analysts can generate robust insights in a fraction of the time it would take a seasoned expert.

We’re also tapping into LLMs for upskilling, particularly in coding. These models assist analysts with queries related to Python, R and data science, making them valuable for teams working remotely or with limited resources. It’s crucial to ensure proper legal agreements to protect sensitive data and prevent it from being used for public training. With the right precautions, these tools significantly boost productivity and skill development.

How do you see the future of this space evolving in the next few years?

In a few years, as AI becomes cheaper and data is better organized, companies will see much faster results. AI will live up to its potential in areas like skew rationalization and planogram recommendations, empowering merchants and streamlining analysts' work. While this future is not yet fully realized, better data will make it achievable.

What advice would you offer to your peers in the industry for achieving success or strategies they should implement?

Partner closely with your business. To succeed, you must meet their needs. It's easy to create impressive algorithms, dashboards or visuals, but they’re meaningless if they can't be implemented due to system or process limitations. Understand your business, its operations and the systems your teams use. While working toward future innovations is important, always deliver value based on what can be achieved today. This approach ensures continued investment and growth for your team