We can add start and end time to get the page load but TC is not actual performance testing tool so we can't say that time is absolutely correct.
Can you please check with "Performance Counter" feature in TC, it may help you.
Tools -> Current Project Properties -> Performance Counter -> Add -> Enter your IP -> select you desire test type
Well, you're doing this on a desktop application and effectively using a stopwatch.
But for the stopwatch to stop, you are checking various object property states to determine when something is complete. And to do this, TestComplete has to check the updated state of object(s) each time you loop. Which can indeed cause delays as it refreshes the object map.
So your times are likely to be artificially high. I would hope not by much, but I imagine if your application is a very busy object model, it could increase the times by more than you'd like as the refresh rate will probably slow down, the more objects it's having to deal with.
How much higher are the reported times vs what you would expect?
