Initially is it a good idea to validate and launch a new concept via a desktop website or an app?

depends on the market, so it's IMPOSSIBLE to advise the correct approach/choice in this question. However, I'd lean towards a desktop/web app for anything that is trying to accomplish 'platform' level mechanics.... most noobs will use the term platform and app interchangeably but there is a huge difference: platforms are designed for OTHER COMPANIES/brands/power-users to BUILD FUNCTIONALITY on top of/extend the use for their specific purposes... TWITTER is a platform. "an app to buy/sell stuff in X new way or Y speciality or Z local/geographic market" is just an app that is trying to say it's cool by calling itself a platform because there's type major types of uses & 'transactions' involved.... a CHATROOM is an app... TWITTER is a platform. Hope that helps.

Generally speaking, I think building a web app for your marketplace concept first can be a great option to launch, then move on to develop a mobile app later on whenever it makes the most sense. Building a web app (vs. a native mobile app) will usually require less time, effort, and money, and if you build it right (responsive, etc.), it can potentially work pretty well across desktop, mobile, and tablet.

That said, there're many other considerations to think about. Here're a couple quick examples:
- target audience & their environment: if you're targeting millennials, mobile may be a lot more important than if your application is typically used by professionals during working hours (in which case web first is a definitive winner in my book).
- hardware needs: do you need to leverage camera / mic / video / sound / accelerometer / GPS a lot? if so, you may be better off building a native mobile app.
- validate: keep in mind you'll have to validate your concept at many different levels (problem, solution, marketing, channels, UVP, cost structure, price & revenue structure, etc). Building a web app can help with lots of this, but it wouldn't be the first move I'd recommend: if you haven't yet validated at least some of your assumptions around problem, solution, and customer segments, I'd suggest you consider validating your idea further before starting to build. And, as a side effect and while you do that, it may also become clearer whether a web app can work or if a mobile app is a must-have.

Just a quick couple thoughts to get you going. I work on this specific topic a lot with my prospects and clients (I build responsive web apps - including marketplaces -, but not mobile apps). I have advised different clients to go with mobile-first and web-first approaches on different occasions, and also advised a couple of them to go back and validate their ideas further before building anything. Let me know if you'd like to setup a call to go over some more specifics to your situation.

This decision depends primarily on your market and the specific needs.

Outside of knowing this key element, my initial reaction is that building a web application that is mobile friendly is a great option to start. If native apps (for various platforms) are desired to continue building your client base, then you can come in well informed from your base web application.