Every decision carries a second, quieter one: how it will be made. Someone has to own it, the options have to be judged against something, and at some point the debate has to stop. If nobody settles those, the decision still gets made, by default.
A meta-decision makes those choices on purpose, before anyone argues about the answer. It covers who decides and who is consulted, the criteria that will separate the options, and the deadline. It also sets how much process the call deserves, which depends largely on whether it's reversible or irreversible.
For a leader, skipping it rarely looks like a mistake. The room agrees that something needs to change, everyone leaves with a sense of progress, and the work starts without anyone owning the outcome.
That's how big decisions stall. With no criteria, every new option reopens the debate, and with no owner, nobody can close it.
The practical tell is the end of the meeting. If the room agreed on a problem but nobody can say who will decide it, by when and against what, the meta-decision hasn't been made.
The board agreed our ERP system had to go, and for months that was the last thing it decided. Nobody said who would choose the replacement, what it had to do or when the choice was due, so Matt, our CTO, started meeting vendors to get things moving. As commercial manager, I wasn't part of it, but the updates kept arriving: another round of vendor meetings, month after month, and still no choice. It took the resignation of the one developer keeping the old system alive to end it, and the leadership team made in a panic the decision it hadn't agreed how to make in calm.