Halima, Good findings can also come from other methods. It depends on your research questions/objectives whether you need to use questionnaire or other methods. Regarding the development of questionnaire, I would suggest you learn how to develop a measuring scale and test its psychometric properties before you actually go for developing it. I do not want to discourage you but if you are new to the field, it is advisable to use other questionnaires validated by experts instead of developing your own. Even checking by experts constitute only face validity which is barely sufficient to be acceptable.
There are many good books available to learn how to develop a measuring scale. For example you can use :
1. Psychometric Methods: Theory into Practice (2017) by Larry R. Price
2. Applied Psychometrics Using SPSS and AMOS (2016) by W. Holmes FinchJason C. Immekus, Brian F. French
To answer your query, I've to say, it's very subjective.
If you're planing to address a issue/problem which perhaps yet to solve by digging in deep, you may wish to develop your own questionnaire based on the problem, question, objective. However, there are certain level of steps to develop all new questionnaire, indeed. Yet you need to consult with expertise from that particular area to validate the questionnaire to avoid error/bias in data gathered.
On the other hand, to bypass those difficulties and possible error, you can use already validated and well used instruments. At this stage we usually define between the concept adopt and adapt.
However, I must say, there is no universal rules in research including expert validation, yet it does makes some sense, but you have to define your own chosen steps. I've attached a link of reading on this issue.
It is certainly part of basic methodology for surveys. You need to find a few experts to check or test the face or content validity of your questionnaire, i.e. whether you ask quesitons as intended or based on the research questions. For a face validity, just ask a couple of experts or colleagues to have a look at it and make some comments on accuracy, wording, etc. Regarding a content validity, you have find at least 5 - 10 experts depending on your work to check it and agree on each item at least 80% of experts. Aside from that, you can check other types of validity, if feasible, plus reliability during a pilot. In the case of a standard quesitonnaire or scale, a full evaluation of psychometric properties is required, i.e. practicality/usability, validity, reliability, and responsiveness.