Tag Archives: Tetlock

The Key to Making Good Decisions

vladislav-babienko-KTpSVEcU0XU-unsplashSorry for the clickbait-y post title. To be honest, I toyed with the idea of having a 15-20 word headline and then decided against it. Of course, there are lots of important factors that go into making good decisions and there’s been plenty of that written about online (including here on this very website). There is one thing, there is one piece of advice, though, that I don’t often come across when it comes to how to make good decisions. In a word — Fermi-ize. Huh? Yeah — Fermi-ize.

A few weeks ago, the superforecasters on the Nonprophets podcast were talking about the possibility that the current POTUS will be impeached. Part of this is because it’s “in the news” and partly because this is one of the forecasting questions that’s currently up on the platform. The framing of the question is important (as is the case with any decision!). The question asks when will POTUS cease to be POTUS (before inauguration day 2021, on inauguration day 2021, or after inauguration day 2021). We’ve got all the important components of a good forecasting question (i.e. falsifiability, etc.).

If you think about the different buckets to that question, you’ll notice that different things would have to happen for the question to resolve in a different bucket. For instance, some sort of untimely death would mean the answer is before inauguration day. So, would impeachment/removal. On inauguration day could mean that POTUS loses his bid for re-election. After inauguration day could mean he’s re-elected. It could also mean that there’s some sort of power struggle with the 2020 election whereby the vote is so close, that we find ourselves with some sort redux on Bush v. Gore (2000).

Circling back to the superforecasters podcast, listening to them (disclosure: I’m a superforecaster) discuss the prospect for impeachment was fascinating. And the main reason is because of how they were fermi-izing the problem. Right, I didn’t explain that earlier. Fermi is one of the scientists responsible for bringing us the nuclear reactor. You may have come across a question like, “How many piano tuners are there in Chicago?” This kind of a question is sometimes referred to as a Fermi problem, whereby you’ll need to breakdown the question into component parts. What’s the population of Chicago. How many people live in each house (on average). How many households have a piano (on average). How often is a piano tuned (on average). Etc. Etc. That’s Fermi-izing.

So, on the episode, listening the superforecasters Fermi-ize the prospect of impeachment — it reminded me of how decisions in the C-Suite unfold. Often times, there are a lot if inbuilt assumptions in a question that might be unbeknownst to all the people sitting around the table. In fact, it’s possible that the question might even be completely wrong. Maybe the question (if we put it back in the political terms of POTUS above) being asked is, will the House vote to impeach POTUS. Many folks think that this is a “slam dunk” that the House will do that. But, what if you’re interested in whether or not POTUS will remain as POTUS. Well, then, you’re probably more interested in whether the Senate will vote to remove POTUS from office. However, even that question might be wrong. For instance, what if the House votes to impeach, and before the trial can be had in the Senate, what if POTUS resigns (!). If you’re focused on the prospects of a Senate trial, you’ll miss the possibility of a resignation. If you’re focused on the trees, you’ll miss the forest.

~

The key to making good decisions is asking good questions. And when it comes to answering good questions, you’ve got to breakdown the question into its component parts.

Outcomes vs. Outputs – The “How’s” of Decision-Making

71GiSvm+a0LRecently, I read (er, re-read?) Phil Tetlock’s Superforecasting: The Art and Science of Prediction. This book came out a couple of years ago (and was co-authored by Dan Gardner, whom I believe is a senior advisor to Prime Minister Trudeau – or was at one point, I’m not sure if he still is). Anyway, the book is excellent and I highly recommend it to anyone who wants expand their understanding of decision-making.

The reason I’m mentioning this book today is because of one of the chapters: The Leader’s Dilemma. The crux of the chapter is a juxtaposition between the central theme of the book – individual superforecasters and their ability to remain steadfast in the face of uncertainty and the fact that leaders need to take action. They can’t waffle in their decision-making – they have to choose. The chapter invokes a German General from the 1800’s, Helmuth von Moltke. To be frank, I haven’t read many books (or seen many documentaries/videos, for that matter) about military leadership and strategy, so this was new to me.

You: Jeremiah, why are you talking to us about German Generals from 200 years ago?

Me: I’m going somewhere with this, I promise, just hang with me for a bit.

So, Moltke. He had a particular way of leading that is often characterized by the German word Auftragstaktik. In English, we know this as mission command. Essentially, this style of leadership boils down to top-down intent, but bottom-up(ish) execution. For example, the leader of the military would say that we need to expand our defenses west, but the leader wouldn’t say how we’re going to do it. Instead, the order makes its way down the hierarchy – absent the how, every time! – until someone would then be executing on that order.

The key here is that there’s no prescription for how to do something when the order is given. The emphasis is on the outcome. Expand our defenses west. Invade that country. Secure that town. OK. So, why is this important? Moltke: “In war, everything is uncertain.” If a General were to give prescriptive orders about how to takeover a bridge, there will absolutely be things that occur on the ground that aren’t accounted for in the General’s orders. Maybe there’s heavy rainfall, which presents a problem for the troops who were told to stake out in the nearby field and wait for the cover of night to execute the plan. Maybe when they arrive, the enemy has three times (!) as many soldiers there to defend the bridge. What will they do? Well, they’ll have to callback to HQ, as everything’s being run through the General.

/It’s at this point that, if I were doing a video of this, I’d splice in a scene from Family Feud where Steve Harvey (or Ray Combs) points to the big board and says, “Show me… waiting for orders from HQ when you’re under heavy fire and have no place to go…”

/Stttttttttrrrrriiiikkkkkkeeee

Yeah, that strategy will not pan out – every time. OK, so we’ve established that in the military, they push decision-making down the hierarchy. How does that relate to us, you might be asking?

Well, how many things do you do at work where you have complete control over how you do them? We need to brief the Director, quick write a briefing note.

/Stttttttttrrrrriiiikkkkkkeeee

What if a briefing note isn’t the best way to brief the Director? What if the Director would rather a quick 2-minute meeting to explain what’s going on, rather than you spending the next 2+ hours crafting the ‘perfect’ briefing note, only to have your manager spend an hour after that re-writing the whole thing? And how does it make you feel after you’ve done all that work, had most of it invalidated by your manager, and then the Director calls you and your manager in anyway because he doesn’t want to read 300 words on the topic, but instead, wants to have a quick chat about it.

~

Or let’s say you’re working in operations – the front lines – where the proverbial ‘government meets the Canadian.’ You’re working with a Canadian who needs a new passport, but is having trouble getting it. Your performance metrics are clear – you’ve already spent 20+ minutes on the phone and if you stay on the phone longer, you know that you’re not going to be able to meet your performance objective because this call will inflate your average time on call.

/Stttttttttrrrrriiiikkkkkkeeee

Clearly, time on call isn’t the best metric to use to evaluate folks who answer calls from Canadians, but the order has already been given from down on high (we need to spend only x-amount of time on the phone with Canadians because it’s part of my performance agreement that our average call-time improves year-over-year).

~

Or how about this – the DM has decided that the department is short on funds so we’re going to reorg two branches – they’re now going to become one branch. *gasp* The ADMs filter the message down to the DGs that they want these directorates axed, the others merged, and some others expanded. The DGs filter that message down to the Directors about which divisions will be axed, merged, and expanded. The Directors filter that message down to the managers about whose role will be shifting, reassigned, or expanded. The managers filter that message down to their teams. And at the end of this exercise no one is happy.

There were no consultations, no engagements, and no empowerment. People were told what they were going to do and how they were going to do it. No one considered that the people they were managing might have ideas about how to do things better. The person filtering the message down was always assumed to be ‘right.’ That’s no way to run a firm and it’s certainly no way to manage change in an organization. Push the “how” of decision-making down the hierarchy. The people you hired are smart (otherwise you wouldn’t have hired them, right?). Let them prove your decision to hire them was a smart one. They can’t wait to do it.

This post originally appeared on GCconnex/GCcollab.