Of course, if you have solved the problem before but think you could do it as a product and sell it, than a lot of the risk and unknowns are removed so you can be way more aggressive.
Personally, I like to market and get firm commitments, plus build up a list of needs as the product is being developed. And as soon as I can I start converting some of those commitments into sales by putting them on the system, I get them paying. This means some might be able to be on it in a month, others might have to wait 3 months while specific features are flushed out.
To me there is real exposure and risk to taking money for a product you can't deliver quickly (~30 days). Tesla is selling refundable reservations which is different and even then they have a material exposure if a large portion of reservations decided they wanted to back out. Also remember, there are banking rules around presales and credit card transactions especially as it relates to refunds, charge backs etc.
To clarify, started building then paused means I did it one Saturday afternoon and never touched the code since then. :)
I used to schedule some of my tweets automatically in a spreadsheet, my wife really wanted it for herself, another friend of hers wanted it, and so I made it slightly more generic: you sign up with google, create a spreadsheet, and connect your twitter account.
Of course it was too primitive, so my wife and her friend aren't using it.
However, from nothing, 1, then 2, then 3 twitter accounts started using it, and they are using it daily, mostly to send out quotes.
So to the original point - I guess blogging about your idea like if you have the tool available to yourself, plus sharing it here & there may help getting some initial interest.
I wanted to test the appetite for retail stores for an application that enables them to sell online, via Chat apps like Facebook Messenger.
Here are the steps I followed. Step1: Put together a couple of videos in Keynote - https://goo.gl/iHzhDg - https://goo.gl/HL847L
Step2: Put it on my phone VLC player
Step3: Walked door-to-door "selling" it. Basically, walk in to a store and say something along the lines of "I am this awesome person who is helping a few businesses sell online via Facebook Messenger. While I cannot show you all the details yet, here's a peek at the product (use the videos to explain a couple of use cases). Would this be something you would pay $XX/month for, to be a part of? If you are interested give me your phone number, I can call you when I let my next batch of customers in"
Step 4: Out of 20 stores I approached, 12 or 13 said they would pay. (If they are asking questions about your product andhow much it costs, that itself is a huge win.) I then discounted that by 80% to predict that even if only 1/5th of the stores that said they would pay, actually paid, i would have got at least 2 paying customers if I had a real product that day.
It actually took me a year from that point, but I am "piloting" with my first customers (1 from that intial walk-in) this week. This what eventually the product looks like. https://goo.gl/ShkkWr
In my opinion, you should definitely sell your product before you start coding. It will not only build a customer base, but tell you IF there is a customer base and make the product you eventually build, better.
If you do it with a video like me, Its OK to fake a few details to make it seem as real as possible. Trust me, its better to rot in hell for "lying" to a potential customer, than explaining to your wife why you spend the savings on building something, nobody wants!
Showing them a mockup or signing them up for email updates at least would be something tangible so they have some piece of mind between the time of purchase and launch.
Tesla did this with the Model S, no? :)
This is a great way to check if there is a paying market for your idea.