There are several different ways to test your FTP. The key is to pick a testing protocol that is consistently repeatable and minimally influenced by outside factors. Let’s face it, few of us have the financial means or the time to be tested in a lab. The test can be run inside on your trainer, or you can pick the same stretch of road outdoors. Clearly, you will need no impediments (stoplights, stop signs, too many turns in the road, etc.) which can skew the test. Also, downhill stretches will skew the test down as well. If you live in an early with long, sustained climbs like here in Boulder, then testing up a climb with moderate grades (2-6% most of the way) is ideal. Ultimately, you want to find a way to allow yourself to get into a solid rhythm which promotes steady output.
For the actual test, while you could ride an hour as hard as possible and use that for your FTP, it’s not a realistic and repeatable test. It’s mentally too taxing, it’s difficult to do the test outdoors logistically and in completing the test solo you’re not going to push yourself as hard as you would in, say, a 40k TT. There are two protocols which tend to be used. The first is the “gold standard” that most coaches default to – 20min and then multiply your average watts by 0.95 to arrive at your FTP. The second test is a 30-minute effort and then taking the average watts at face value as your FTP.
After testing many athletes and going through countless FTP tests myself, here’s my take on it. Either shorten the 20min test to 15min and still multiply by 0.95, or complete the 30min test. Riding longer seems to make more sense and allow for a more accurate test result because you have to avoid starting out too fast and really focus on settling in to a sustainable effort. On the flip side, shortening the 20min test by 5min allows for a more maximal effort that then needs to be handicapped by 5%. My experience is that 20min is more of a “no man’s land” duration – too long for a true maximal effort but too short to allow for settling in. My 2 cents.
Whichever way you decide to go about it, make sure you decide on a protocol that you can get excited to replicate over time so that you are comparing apples-to-apples one test to the next. Also, frequency of testing – every 6-8 weeks is good enough. Training zones are minimally impacted by a handful of watts, so no need to test every month. Lastly, conduct the test at either the very end of a recovery week or the very beginning of a training cycle after a recovery week. This ensures you are rested both mentally and physically, and ready to give the test the effort it requires.