Usability doesn't just happen. Usability must be consciously incorporated into a product from the start. And the start begins at product conception.
Computer and software firms are primarily technical firms. When they start to think about a new product, they almost always think of it in terms of new capability they will be designing into the product. From the start, the product is conceived of in terms of new features. But what is critical for product success is not what features the product has but what the user is able to do with the product. If the user cannot use a feature, then that feature doesn't exist!
This means that the first step to usability is to determine what the user wants to do and should be able to do with the product. This change in viewpoint is needed to focus the design of proposed features on the user. It should also raise the question of how valuable a given feature is to the user. There is no sense pouring large amounts of engineering budget into developing a feature that users don't value while ignoring those that provide significant benefit to the user.
However, it doesn't take long to discover that the value of a feature is contextual and dependent on when and how a user will access and use the feature. For example, a VCR which allows a user to start recording with the simple press of a single button on the remote is likely to result in a number inadvertent over-recordings - especially if the Record button is close to the Play button. While the ability to record would seem essential to a VCR, an inappropriate implementation such as this might make the user regret having the feature if not the VCR itself.
Usability starts at the beginning of product conceptualization by getting to know who the users are, what their needs are, what the context of use is, and how they will be using the product. This is done through ethnographic studies, market research (surveys, focus groups), competitive analysis, and contextual user studies and analysis.