I'm currently studying interpolation spaces, but I'm struggling to understand some of the intuition behind real interpolation.
There are two equivalent methods to define the real interpolation functor, namely the J-method and the K-method, and both of them use the $L^p((0,\infty),\mathcal{B}((0,\infty)),\frac{1}{t}dt;\mathbb{K})$ spaces to define the real interpolation spaces (where $\mathbb{K} = \mathbb{R}$ or $\mathbb{K} = \mathbb{C}$).
I have little problem understanding the definitions and proofs on a formal level (after working them out, at least), but I really don't get how someone would think of using this particular measure instead of the Lebesgue measure $\lambda$. I understand that the measure $\mu$ defined by $d\mu := \frac{1}{t}d\lambda$ is the unique Haar measure on $((0,\infty),\cdot)$ and therefore has useful properties similar to the Lebesgue measure, but that doesn't seem like a real motivation to me.
Can someone provide some intuition behind the real interpolation functor, preferably the K-method?