I took a startup course in the MBA program at one of HBS/GSB/Wharton/Kellogg that was taught by a (in my estimation) great VC. Arguably the most insightful class I've ever taken. I'll give two sets of insights: anecdotes and principles.
(A1) The founders of a successful custom apparel startup didn't get even close to 100 customers until after they had tried two different distribution channels. First, they opened a store and that didn't work. Second, they tried to sell on-line and that didn't work. Finally, with implosion imminent, they tried a hybrid model where you get measured in the store, but every purchase is made on-line and then delivered. It worked.
(A2) Co-founder of one of the publicly-traded food delivery companies told us how he started out. Was hungry, realized he couldn't order online, so he sits down and codes an initial web page (as well as an API). Asks a friend of a friend that runs a restaurant to let him list the restaurant on his new site. Guy gives him $140 for one month.
Figures he'll just knock-on doors to add restaurants, but managers kick him out with no meeting. Realizes if he walks in the back door (alley) and pretends to have a meeting with a manager he might get somewhere. So, he walks through the kitchens of 100's of restaurants and they ask him who he is there to see, tells them he has a meeting with the manager. 50% of the time he gets kicked-out, the rest of the time he gets restaurants to agree to being listed.
(A3) My VC professor's fourth business was a bridal gown website he purchased with his co-founder. They didn't use paid SEO. Rather, they came up with creative hacks to drive traffic to the site. One of the better ones:
Pay writers in India to compose short articles about stuff happening in the bridal space. The writers read the first articles to be published that morning (UK time) in the BBC, etc., then write up the same stuff, and publish it on their bridal gown page right as (US time) the same stories were hitting presses at the NYT, etc.
Using such hacks, they got their site to number two on google search results without paid SEO.
-----
(P1) "Half of the difficulty is going from zero customers to one customer. Everything after that is the other half."
(P2) Every startup is a science experiment. Come up with a hypothesis, test it, and see what worked and what didn't. Rinse, repeat.
Following a structured, hypothesis-driven process is key. Literally, write in a journal what your hypothesis/assumption is, how you are going to test it, and what you hope to learn from the test. Next week, write-out what your hypothesis is, what you need to do this week to make further progress on validating/invalidating it, what progress you made the week before in validating/invalidating it, and what your plan for the following week is. Continue writing out these four things every single week to stay organized.
(P3) If you feel extremely strapped for cash, view this as a blessing in disguise. There is some creative hack that will allow you to validate/invalidate your hypothesis or set of assumptions - you just haven't found it yet. Running solid, structured experiments (i.e. process) is most important. Love process.
-----
Unless you have blown way past 100 customers, the sciency/experimental aspect of a startup doesn't change. And even inside unicorns, management is (hopefully) still ultimately just running a never ending series of experiments.