Huh. I think this is often taught. Indeed, I think it's at the core of the LS approach. The build-measure-learn loop is the heart of the process, meaning that whatever ideas you have, you test them as soon as possible with customer
behavior.
As an example, I mentored at Lean Startup Machine. We'd have them form up into teams around ideas. The first thing they'd ask for was verbal feedback. But they were quickly supposed to move to increasing levels of demonstration of commitment.
One team, for example, had the notion that restaurants needed an easier way to get insight into customer experiences. They went out and found restaurant managers, and first confirmed that they had the hypothesized problem. But they quickly moved beyond words, asking if the managers would be willing to let them trial the system in the restaurant. Since enough of them said yes, they build something simple and put those little printed cardboard triangles on a bunch of tables. They then hacked together reports for the managers based on the data collected. The managers liked the reports. They then asked the managers how many would be willing to pay, right then, some modest monthly charge for ongoing use of the system.
If I recall rightly, this team discovered that when actual money was on the line, the reports were just a nice-to-have. In that situation, it's still reasonable to persevere, as you may have hypotheses about what really will get them to cross the line. But it's also reasonable to kill the idea.
There's a big difference, though, between what is taught and what people do. I was an early adopter of both the Agile movement (XP in particular) and Lean Startup approaches, and in both cases I'm kinda horrified by what it turned into in practice. With the Agile movement, it was quickly captured by companies who wanted to feel like they were doing something different while still retaining most of their top-down pathologies and high-debt code bases.
With the Lean Startup approach it's harder to say how it turned out, as early-stage startups rarely produce reports like this. But my general impression is that, as with Agile, people pick the parts that seem fun and easy, and avoid the hard parts. I think this was made worse by the flood of VC money, especially around Uber-for-X startups. Who wants a method that will kill your probably-bad idea early if you can get tens or hundreds of millions? The bad idea might work out or you might thrash your way to a good one. And if it still failed? Well, now you're a seasoned founder of a venture-backed startup. By some measures, although not mine, that's way better than having a quiet, cheap failure.