Imagine you're working with a retail company's sales data, and the dataset contains transaction records with varying units of measurement for products (e.g., pounds, kilograms, pieces). How would you normalize the data to ensure consistency for analysis and reporting?