A dual-class structure creates two classes of common stock that are identical except for voting. Founders hold Class B common stock with, say, ten votes a share, and everyone else holds Class A with one. Section 151 lets a Delaware corporation create classes with different voting rights in its charter.
The goal is control. Each financing dilutes the founders' economic stake, but with high-vote shares they can keep a majority of the votes long after they own a minority of the company. Several large technology companies went public this way.
Most startups do not adopt it at formation. Seed and Series A investors generally accept that founders control the company early because founders own most of it, but many object to locking that control in permanently, and some funds will not invest in companies that do. It also buys less than it seems to. In a venture-backed company most important decisions belong to the board, whose seats are negotiated in each financing, and investors' protective provisions constrain the company whatever the vote count. Two classes of common mean class votes on some matters, longer documents and more diligence questions. And investors who accept dual-class stock usually require the high-vote shares to convert to ordinary shares when a founder leaves, sells or dies, or after a set number of years, which removes much of the benefit.
Founders who expect to keep a large stake, raise from investors comfortable with the structure and stay in control through an IPO sometimes adopt it before their first priced round, when the founders' votes alone are enough. Later it needs the investors' consent. If you want it, decide early and use a lawyer. Corply's standard charter has a single class of common stock.