The first point is often totally wrong. I worked in a place where the sales team sold the product hard before much of it was built and before it was reliable.

It turned out to be way harder to build what customers actually wanted than anyone knew ahead of time. Nobody could foresee it. The only way we could have known we were egregiously over-promising in the sales pitches was to just actually start building it first, see how hard it would be and what the unexpected sources of engineering difficulty would be, and then scope our sales vision or pitches accordingly.

That particular product team failed badly and the product line was closed down. We started selling before building, and only later realized that when you do that, you’re talking out of your ass and you are doing a huge disservice to prospective clients who buy into your sales pitch just to be let down later when you cannot execute on the implementation for reasons that could have been known if only you had invested in building things first.

The same issue can also play out with costs: maybe you technically can build what you sold, but because you over-promised in the sales pitches, you end up in a situation where to be able to offer what was actually sold, the engineering costs force it to be intrinsically unprofitable, while had you known the cost projections from building first, you might have been able to scope the sales pitch to only a profitable set of features.

It’s way easier to throw away / rebuild something later when sales feedback indicates you should pivot than it is to build nothing at all, over-promise, and try to keep clients who end up unhappy that you misled them.

The cost of building first so you have adequate information about what it would require to profitably support selling a certain product or feature set is known as _investment_.