All derivatives are financial contracts, and in these contracts, there are far more features that can be agreed on than a simple right to buy or to sell. Complex payout structures can be engineered based on what-if scenarios; thus, the final payout of an exotic contract can be dependent on a whole set of circumstances. Often, even the path of the underlying has a serious influence on the final payout. Compared to these derivatives, the good old call and put options were soon seen simple, earning them a not too impressive nickname: plain vanilla.
Vanilla call and put options are like plain vanilla ice-creams, the simplest possible ice-cream without any fancy optional toppings. The expression "plain vanilla" is so strongly embedded in finance that it is even used in the bond market, where a vanilla bond is the simplest possible coupon-paying bond.
Any option that has some extras over the basic plain vanilla options belong to a very numerous group called exotic options. Exotic options are popular because sell-side bankers are in fierce competition to offer tailor-made products for the clients. Another reason behind the fact that exotics are widely spread is that, interestingly enough, most of the time, quoting a price on an exotic structure is not a much more difficult task for market makers than quoting vanilla prices.
Exotic or not, there is one intrinsic feature that is always the same in every derivative product, that is, it is a function of other instruments, hence the name derivative. Thus, the price of a derivative is not independently developed as the outcome of a direct supply and demand; rather, it is given as an estimated construction cost. For example, the one month forward dollar price of a euro is highly dependent on the spot dollar price of the euro; the forward price is just the function of the spot price (and the interest rates).
If exactly the same benefits that are granted by holding a derivative can be constructed by a trading strategy that involves less complex instruments, then the derivative can be replicated. Derivatives are not like unique paintings; the forgery of a derivative has the very same value, while replicas are as good as the original. By using the no-arbitrage argument, Black and Scholes (1973) and Merton (1973) showed that the price of a derivative should be equal to the expected sum of expenses that arise during the proper implementation of the dynamic replication strategy. Taleb (1997) extensively describes that implementing a proper replication strategy under real market circumstances could often be really tricky.