Every piece of work you do is a prediction. The amount you find out how wrong you are scales with time you fuck around
Disproof
Release and test stuff maybe?
Consequences
What is really counter-intuitive is that the incremental amount you can ‘perfect’ something (or try to anticipate and ‘pre adjust’ for a potential correction) tends to diminish per day, but that error tends to grow per day.
i.e. There’s a window for validation and correction. If you overshoot it, the pain is going to ramp much faster than you can anticipate and correct for.
e.g. spend 2 weeks building the functionality, then spend 2 more weeks polishing the CSS, but the problem was actually the information architecture in the first place. Ouch
In extreme cases, error feedback loops can happen where you actually get more wrong over time
Contagion effects where entire other departments gear up around an expected state of affairs (e.g. prepping to internationalise something that might have to change drastically after its release)
Releasing something actually bad after a long time, instead of something imperfect the can be fed back on after a short time, almost categorically loses more users
Spend enough time backtracking or scrambling to fix something you should have released weeks ago and you risk running out of money, customers, goodwill, etc.
Causes
Almost always caused by some someone somewhere in the chain being scared to release something ‘unpolished’ for fear of blame/ridicule/personal liability
If this fear remains, you’ll never change
Most other times, someone doesn't want to 'tip their hand too early' and get outcompeted
Approaches
Correct the thinking and fix the fear of the unpolished with an even bigger fear of the consequences
Every prediction will have some error in it. You find this quantity out by contacting reality (both in test and in launch). Once you know the error, you can correct it.
What should be obvious is that as time between prediction and contact with reality grows, so does that error. This means you’re almost always buying yourself a bigger correction the longer you go on uncorrected
If you’re working on something that just takes a long time, and you have money, and people will entrust you with years of their time, go ahead. But most of us don’t have that, and even if we did there comes a point where you’re leaving time on the table.