SAFE is definitely the best way to go if you're pre-revenue. Post-revenue, you can use exit multiples on top-line revenue to get close. I'd recommend having the investor suggest it and use a market-valuation if you don't want to set it yourself. It's difficult really unless you have record of valuations of other companies who are in the same industry when they raised a similar round or had similar milestones completed. Keep in mind a valuation isn't really a precise science but it's a measure of what the market is willing to pay to buy the equity you are selling.

First thing I would say is read The Lean Startup by Eric Ries. He goes through the process of validating an idea very clearly in the book. Bottom line, you should find advertisers and validate that they are willing to pay for what you want to build, then you need to find your users and make sure that it is a system of value to them. There's a whole host of product development firms and employees that you could hire, but ultimately, it doesn't make sense to build something until you've got your users saying they'll use the platform and your customers saying they'll pay you (preferably in writing).

There's a lot of options, but a form that's getting popularity recently is the SAFE (similar to a convertible note) which was popularized by its use in Y-Combinator deals. It allows you to set a valuation cap on the round for those early investors and set a discount for their early participation on whatever the valuation of your round ends up being. It's a great way to postpone the valuation and terms to a later date, especially if you still need to prove out your business model.