You say you have "a startup idea (a web service), which I consider profitable in general", but what does that really mean? Have you done market research, or is

You say you have "a startup idea (a web service), which I consider profitable in general", but what does that really mean? Have you done market research, or is this based on your gut feeling? If the latter, then you should do some market analysis, and one way to do this is a Dry Test. (It's not the only way, of course.) Except for dry testing, I don't see any reason to make a "coming soon" website, if you haven't got an MVP in place. Once you have an MVP, by all means, launch. The sooner you get real customers using the service, the quicker you can get feedback, and iterate.