A correlation matrix allows you to quickly identify interesting patterns in your data, empowering you to see how your data relates to one another. Do metrics move together, opposite one another, or totally unrelated? Two important actions can result: if you take actions emphasize one of the features in your model, you know how it's likely to impact other aspects of the business. For example, if I increase marketing spend, it will have adverse or reinforcing effects on other KPIs in my business?
Second, metrics that move together and are highly correlated are likely redundant and merit possible exclusion from the model. See section on Correlation is not Causation.