How to measure Windows softwares' performance to make it exactly match user experience?

See more:

C++

Windows

User performance experience focuses on what is the full cycle time from when the user clicks a button till a new dialog is ready for interaction.
How can I measure Windows softwares' performance exactly to match user experience?
How can I capture the Start Time("from Click")?
How can I capture the End Time("Every thing is done and the dialog becomes interactive")?
The most important thing, I want to make this kind of tool to monitor the other processes'(mostly no pdb.) performance.

Please describe your solution by using C++、Windows、Process、Thread、Message Queue Or any other professional terms. Thx very much!