Having done both, and everything in between… I’m continuing to move steadily away from binary implementation decisions.

You spend money, it gets done quicker (sometimes, but that’s the intent) but if it’s not the right thing, it’s a waste.

You spend time, you have time to adapt and adjust, but you can be too late.

I’m moving towards smaller, cheaper, and faster incremental interventions. Accrete successes, scrape off failures.