Three Things To Put Off As Long As Possible For Your Startup

On this point, the advice given is great is you are a sole founder but even then may create tax risks if you time your entity setup to match the time of your first funding (e.g., Founder A gets 90% for IP assignment + Investor X gets 10% for $100K = Risk of $900K taxable income to Founder). This tax risk is normally manageable but it also is most easily avoided or minimized if even a sole founder sets up an early shell (even an online instacorp) to give himself a buffer between first founder grant and first investment.

With a founding team, it is normally a mistake, in my judgment, to defer entity setup too long as people invest time and effort and develop IP for the project. It can and often does all work fine as long as the co-founders in such a case continue to cooperate right up to the point where they chose to document their relationship but they are at risk of, e.g., "I was promised 25% of the company" type of claims or "I developed key IP that belongs to me" type assertions if they have not defined stock grants with strings on them (restricted stock) and have not entered into work-for-hire contracts should they have a bust-up or serious disagreement of some kind prior to documenting things.

The practical risks of a disagreement may be low but my point is that, when you have a founding team especially, it is a mistake to have an axiomatic rule along the lines of "don't incorporate until you are working with Other People's Money." The issue should be consciously thought through by any founding team and the cost/benefit considered before deciding this one way or the other.

A more detailed analysis is found here: "When should I set up an entity for my startup business" (http://grellas.com/faq_business_startup_007.html).

It's very, very tempting to "experiment" and "test" various paid customer acquisition channels. You start with AdWords, then do some cheapo banner ads on a couple of ad networks, and soon you're hooked on some hard affiliate stuff and paying thousands of dollars a month to an Outsourced Program Manager"

Humorously worded, but I can tell you from loads of experience, this is flat wrong. Your LTV for customers you acquire from ads can be vastly different (in either direction) from ones you gain organically. I've had apps where it started out 3x CPA, then ended up 1/3x as the law of big numbers came into play and we had reached everyone we could target well.

Experimentation is how you grow and find your place in the market. From the start you should be throwing small chunks in different places, testing the signup flows of the people who come through them, looking for your highest ROI and most scalable method of customer acquisition.

If you don't have the self-discipline to wait until you're showing sustainable profits to throw the big budget in, and carefully monitor it throughout, then you shouldn't be running a startup. You should be waiting tables at Applebee's.