If breathing into a paper bag doesn't help, try focusing.

If you can't focus, try pin pointing and completing smaller projects, rather than taking on one or several large ones at once. Usually, a smaller project can be built into a larger project. Most things are expandable.

You'll learn how to turn smaller projects into bigger ones only if you first complete the small project as at least minimally working. Usually big projects end up not getting past the testing/experimental stage due to a feeling of overwhelmingness, anxiousness, design questions. Smaller projects may be launch as test projects, and then be reliably working before you know it. If you first finish a small project which can later be built into the big project, it can save the big project.

With regards to software projects: building a strong small base program which is functional can help. Build the base first, then the features. If you have problems with building features before the base, you may have a creative syndrome where you only wish to work on interesting and fun things. Find fun things to work on that are in fact part of the base or core of the project. Usually there is always something part of the core that is fun to work on, just that you get side tracked with features which are fun.

With regards to web sites: design the site so that it can be expanded as you go, not done all at once. Design it so that it can be broken but still working.. such as this wiki in some ways. The brokenness of a website should be invisible to the user, if possible. If the website is broken but works properly, it's actually just missing features. It isn't actually broken.

If a website is not at least partially broken in some ways, there is something wrong with the web designer. People who pay for websites to be "complete" are kidding themselves. Those 6 or 10 page web sites in "perfect working order" never make it far on the internet. Google for example has many broken parts in it's database, errors in result counts, and many more issues. But it's made invisible to the user. This is key. Large successful websites are never "complete" but they have a strong base that is at least functioning reliably.

Websites can afford more errors than in software projects, but it's better if the error is invisible to the user and of low priority. Websites are generally large sources of information whereas software is more a precise tool for a specific job. On a website, if you have 5000 information pages up and 100 are down, it's not a big deal. But it's better if that's invisible to the user.. i.e. you may show a count of 5000 pages available, but only display 4900 to the user.

Breathing into a paper bag means to settle down, focus, concntrate, prioritize, and succeed.