We tend to focus a lot of energy on making “good” decisions. But I think it’s more important to think about making the right type of decision for the problem you are facing. 

Most “bad” decision making isn’t bad in a generic way, but “bad” in the sense that it is contextually incorrect. Deciding about which type of jam you should buy at the grocery store is qualitatively different decision than deciding whether to buy a competitor. Buying the strawberry jam because your toddler points at it is a reasonable basis on which to proceed with your purchase. Buying another company because your toddler likes their logo probably isn’t the most sensible move.

Stripe CEO Patrick Collison once remarked that he thought about decisions differently based on their magnitude (how important a decision it is) and reversibility (how hard it is to fix a wrong decision). This converts nicely into a helpful 2×2 I find myself referencing a lot.

Credit to R_Ganesh for the original version of this 2×2

Low Magnitude/Easy to Reverse: Use Your Gut/Coin Flip to Decide

This is probably the easiest type of decision to make but often gets overcomplicated.

How to Decide: Go with your 2 second gut feeling or whatever your toddler likes. Otherwise, flip a coin, it doesn’t matter.

Common Failure Mode: Spending too much time and resources on it. E.g. Having a 1 hour call to debate your company’s exact newsletter formatting. It doesn’t matter, make a decision and move on. If you find yourself dealing with someone that has strong feelings on every detail like this, just let them do it even if it’s suboptimal. It’s not worth the time to debate.

Examples: What kind of Jelly to get for PB&J sandwiches, how to format your marketing newsletter

Hard to Reverse/Low Magnitude: Use Heuristics to Decide

How to Decide: One of the challenging categories of decisions is relatively low magnitude ones but ones that you have to make often. Decision fatigue can set in and your energy gets sucked on relatively inconsequential decisions. Irreversible/low magnitude decisions don’t matter individually but can ultimately matter cumulatively. If there are consistent types of irreversible but low magnitude decisions, try to use heuristics (AKA rules of thumb).

Common Failure Mode: Spending too much time on one-off decisions and not establishing heuristics for repeated decisions.

Example: A customer requests a refund 1 day after the refund policy expires. Does this one decision really matter a ton for the company? Probably not, but it’s a question that is likely to come up repeatedly and so you need to come up with a heuristic that people can easily follow.

Good company culture has heuristics that allow people in the company to make these decisions without needing to spend substantial resources. Amazon focused on customer-centricity most famously which leads to them tending to err on the side of being overly generous. From my essay on company culture:

Beyond the table stakes level, a good culture is one that is optimized to the business model and competitive environment of the company. Amazon’s culture is famously “customer-centric.” This isn’t because they are particularly nice. It is because Amazon’s business model has a strong, positive feedback loop for making customers happy.

More customers mean lower fixed costs per unit sold for Amazon and more new third-party sellers. This in turn drives a bigger selection at lower prices which brings in more customers and starts the whole virtuous cycle over again.

 

Image Source: Futureblind

The nature of the business means losing a little bit of money upfront to make it a good experience is worth it for Amazon’s long-term value. So having a culture where people at all levels err on the side of providing a good customer experience was not done out of generosity, but a strategic business choice.

A business in a declining industry with small margins and a low number of repeat customers will require a different culture of dealing with customers to survive. A customer service representative there might have to be polite but firm about enforcing their stricter return policy.”

Figure out what the recurring types of hard to reverse but low magnitude decisions are for you and come up with heuristics to deal with them. Another simple example is that most diets which people successfully adhere to have fairly simple rules. Paleo is “no grains, sugar and legumes.” Vegetarian is “no meat.” If you’re following a vegetarian diet because you’re concerned about animal cruelty, then you could enter into some complex calculation about how the animal was raised and living conditions to determine whether you felt ok about it. Functionally, this is really hard (you’d have to interview the chef every time you went out to dinner) so just going with “no meat” is a computationally frugal and effective approach.

High Magnitude/Hard to Reverse: Measure twice, cut once.

How to Decide: Your decision process should be robust rather than efficient. Gather lots of data, talk to a bunch of people, and be able to explain the idea maze for why similar decision have gone well or poorly. Stress test your decision against smart people with relevant experience. It’s appropriate for this to consume a lot of time and money.

Common Failure Mode: Entrepreneurs that are good at ‘moving fast and breaking things’ are often able to grow quickly their first few years in business  and get it to a meaningful level of success but they then start to mismanage the more mature business and make decisions too quickly that ultimately are high consequential and hard to reverse.

Example: Major capital expenditure (CapEX) decisions like building a new factory or investing in new software infrastructure (or failing to do so when you should).

A great example is Zenefits. Zenefits helps startups and small businesses find insurance quotes and manage employee benefits in one place.

It was, at one point, the fastest growing software company in history until it collapsed with large layoffs and was ultimately sold off.

In a post-mortem founder Parker Conrad explained what went wrong:

Fundamentally, my view of what went wrong with Zenefits is that I made a decision very early on that we were gonna scale the business faster than our engineering team could really support it.

And so, we thought, look, what we’re gonna do is not everything’s in place yet, not everything’s fully automated.

Companies are doing all this admin work; we’re just gonna do it for them. And it was a classic YC do things that don’t scale, which is absolutely the right way to start a company. The corollary to that, though, is eventually, you have to scale them.
…
So, what happened is we scaled out this manual process. We had a ton of people doing things manually behind the scenes, and that led to a series of problems, and one of them is it actually made it a lot harder to automate things, because once things are being done manually at scale, it’s incredibly difficult to go back and automate it.

The program’s too complex, it’s too hard to build software to replace what people are doing.

And the second thing was that, when we were doing things manually, there’s always an error rate around that.”

When the company was small and it had fairly few customers, the biggest risk was “will anyone pay us for this” and so just doing things manually at a small scale makes a lot of sense.

However, what was a correct decision at a smaller scale (to do thing manually) became an incorrect decision because as the company grew. That choice moved from a “reversible/high magnitude” one into an “irreversible/high magnitude” one.

They needed to reverse that decision when the company was small and it was still easy to do so, but didn’t. This eventually compounded and led to major problems at the company ending with layoffs and being fined $7 million by California’s insurance regulator for allegedly disobeying insurance laws.

Easy to Reverse/High Magnitude

How to Decide: Of the four quadrants, this is the most common one I see people fail at. It’s very important so you feel like you should take a long time with it, but it’s also easily reversible so you should make a decision using the 70% Rule: make a decision when you are 70% sure and just be good at course correcting when you’re wrong.

From a Jeff Bezos Amazon Shareholder Letter:

Many decisions are reversible, two-way doors. Those decisions can use a light-weight process. For those, so what if you’re wrong?
…
most decisions should probably be made with somewhere around 70% of the information you wish you had. If you wait for 90%, in most cases, you’re probably being slow. Plus, either way, you need to be good at quickly recognizing and correcting bad decisions. If you’re good at course correcting, being wrong may be less costly than you think, whereas being slow is going to be expensive for sure.”

Common Failure Mode: Moving too slow and getting outflanked by the competition. You spend too much time making the decision, perseverating over every detail and end up becoming irrelevant. Often this happens when people that worked at big companies move to a smaller company. Decisions that are hard to reverse with 10,000 people are often easy to reverse with 10 people.

Example: Most product launches fall in this category. For example, I launched the first version of my book, The End of Jobs, with probably 100 typos in it. It could probably have been cut down another fifty pages. It was 70% as good as it could have been.

It took me nine months from writing the first word to having the book published and for sale. As you get further along, each incremental improvement take a lot longer. If it took me 9 months to get to a 70% draft, it probably would have taken 18 months to get it to 95% or 27 months to get it to 99%.

Why not take the time to get it to 95% or 99%? It was a digital product so editing the book just meant editing a document on my computer and re-uploading it to Amazon.

Last Updated on March 1, 2023 by Taylor Pearson