Seal hydrogen and iodine vapour in a flask at 430 °C. They react to make hydrogen iodide, H2+I2⇌2HI, but HI also breaks back down into H2 and I2. At first the forward reaction wins; as HI builds up the reverse speeds up, until the two rates match and the concentrations stop changing. Measure them then and compute [H2][I2][HI]2: you get about 54. Start with different amounts, even with pure HI, and the concentrations settle at different values, yet this ratio comes out near 54 every time at that temperature. That fixed ratio is the equilibrium constant.
The exponents come from the balanced equation. Two HI molecules form in each forward event, so [HI] is squared, and doubling [HI] makes the top four times bigger. Pure solids and pure liquids are left out: a lump of CaCO3 has the same concentration of itself whether it weighs 1 g or 100 g, because its density does not change, so the size of the lump cannot shift the balance.
The same fraction built from a mixture that has not yet settled is the reaction quotient Q, not K. If Q is smaller than K, the mixture has too little product and moves forward; if Q is larger, it moves back. Adding more H2 changes Q and makes the system shift, but once it settles the ratio is back at 54. Only a change of temperature changes K itself.
Think of it like this. Picture two rooms joined by a doorway, with people wandering both ways. Each person has a fixed chance per minute of drifting through, so the crowds settle at a steady ratio. Push extra people into one room and traffic out of it grows until that ratio returns. Temperature changes the chances, and so the ratio. Coefficients turn the ratio into powers, which the rooms do not show.
Core idea. At a given temperature, every equilibrium mixture of a reaction gives the same products-over-reactants ratio, with each concentration raised to its coefficient, so that one number K predicts how far the reaction goes and which way any other mixture will move.