I would use split lap timer in one hand and download the Android application with the second. Update a few times, and then calculate the average. This should give a good estimate of the start time. Do not joke .
Does your web application affect Javascript? Or is it server-side intensive, or perhaps both? Network latencies may vary, so it’s best not to take this into account when profiling. Instead, profile the server side separately, which will give you a clear assessment of how quickly the response is generated. Android uses Webkit, so the web inspector should give you a pretty good idea of how Javascript will work on your Android device. You can also check the v8 test suite on your device.. Like these test suites, you can write your own performance tests specific to your web application. There will be many things that can affect those benchmarks that are out of your control, so keep this in mind when writing these tests.