Timing a solve is useful only if you never let it change the move you choose.
Two players can time the same ladder and learn completely different things. One learns that the solve took ninety seconds. The other learns that the first thirty seconds were scanning and the next forty were spent repeating one dead end. Only the second player learned something worth keeping.
The rule for useful timing is that the timer must never influence the move. The moment you hurry because the number is running, the measurement is corrupted. Practise ignoring it: set the timer, put the phone face down, and play the board as if no one is watching. The number is a by-product, not a coach.
With that discipline, timing reveals the shape of your solve. Fast openings and slow middles point to planning weaknesses. Fast middles and slow endgames point to a thin ending vocabulary. Uniform slowness usually means the board is hard, which is fine, and uniform speed with wrong answers means you are moving quickly through decisions you have not made.
The most useful version of this is comparing the same player over a week rather than one board. Times calm down when the format is familiar, and the stable number is far more informative than any individual solve. Visit the site on a steady day and time three boards in a row. Then unlimited rounds give you a supply that makes the weekly comparison possible instead of anecdotal.
If timing starts to feel like pressure, stop timing for a week. The point is information, and information that distorts the thing it measures is not worth having. A player who can time without hurrying gets the best of both: the same relaxed move choices and a number that actually describes them. A single slow solve is a single slow solve. The shape of the week is the data, and it needs at least four boards to say anything at all.