First, didn't realize this was Rand Fishkin writing it. He knows his stuff. He also has another linked article in there that is more prescriptive on incrementality measurement being all that matters and how to think about it at a high level and I completely agree it is all the matters.
https://sparktoro.com/blog/how-to-measure-hard-to-measure-ma...
That in turn links to another respected analytics leader, Avinash Kaushik:
https://www.kaushik.net/avinash/marketing-analytics-attribut...
That is also a must read.
The bottom line is what used to work no longer does, and marketers (and finance and leadership) need to get used to having less fidelity and availability than they were used to. It also means marketing teams who are thrashy trying a new low impact tactic every week instead of constructing experiments likely to deliver statistically significant results on incremental lift are going to be spinning their wheels and wasting dollars. And that's something to watch out for.
Unfortunately, setting up proper experiments, controlling for bias, getting clean data, etc. are material challenges that require skills, scale (to get a read on bottom of funnel especially), and resources/budget.
There aren't great options out there for that now if you're a smaller or mid size company that I'm aware of, though if anyone is aware of them I'd love to have them on my radar.