Typical forms are far out-of-the-money index puts, long volatility exposure, or a small allocation to instruments that rise in a panic. The defining feature is a payoff profile that is negative in almost every period and hugely positive in a rare one.
Price it like insurance, because it is. A tail hedge costing 0.5-1.5% of the portfolio per year needs a crash roughly every five to ten years to break even on its own, and the timing of that crash decides everything. The honest question is not whether it is profitable in expectation - usually it is not, after costs - but whether it lets you carry more exposure elsewhere or survive a forced-selling cascade without forced-liquidation.
The common failure is behavioural: buying protection after the drop, when implied-volatility has already tripled, and abandoning it after three quiet years, immediately before it would have paid. If you cannot hold it through the boring decade, do not start.
Related: protective-put, scenario-analysis, black-swan, hedge-ratio