Don't try to make the RIGHT decisions

You should instead make decisions right.

About one year ago, a friend asked for my advice on what technology he should use for his startup app idea.
He was hesitating between 2 technologies: A and B.

Two weeks ago he told me "yeah man I started with A but it turned out bad as I can't do feature XYZ easily, we need to start again with B or whatever, should've done that from the start. "

I mean, he is technically not wrong, but he had no way one year before to know that, and he doesn't even know if going B would've been a really good choice or not.

2 problems for the price of one.

There are two faces to this problem.
You have the usual "overanalyzing" problem, where we take too much time to make a decision or gather way too much data, and either go with a "default" one looking for so-called "safety" or go with a rash decision because it's the only one.

But we have also the "under-committing" problem, if the decision-making process isn't clear or there is doubt, people don't commit to the decision, and it goes back to the table every two weeks.

We're gonna talk about both those problems, and how to make any decision and turn it into the "right decision"

classic people trying to overanalyze even the decision-making process

1- Overanalysing

If you ever have to make a decision, especially if you think it's important, you can find yourself trying to make the most "informed decision" possible.
A common pitfall is getting way too many data points, too many pros, too many cons, and just too many variables to decide with a single brain.

Your boss is of course waiting for you to take responsibility, you don't want to get blamed for a bad choice, so you proceed with "caution" and always need more data.

Add to that multiple stakeholders and decision-makers and you have what the CIA calls the "Top 10 ways to sabotage an organisation"

Even the CIA agrees about the uselessness of large committees

Tools to solve

It's basically a classic "exploit vs explore trade-off" problem.
Exploiting without exploring is just taking an "uninformed decision" (whether a rash decision or a "safe decision")
And just exploring without exploiting, is just overanalysing.

But some really smart people already solved this problem in several ways, in multiple kinds of situations

If you can reallocate your resources, and you don't already know the payoff (the infamous ROI), you're best off studying the Multi-Arm Bandit Theory

1-1) Multi-Arm Bandit

A multi-arm bandit is a situation where a decision-maker must allocate resources (such as time, money, or attention) among multiple options (or "arms") with unknown payoff distributions

A visual explanation of the multi-arm bandit problem - Source

The Gittins index is a mathematical heuristic that considers both the current performance of a choice and the potential future performance based on the uncertainty of the data.
I don't want to bore you with this, you can find online how to compute it (or ask GPT tbh), for now on, for simplicity I'll refer to it as "The math thingy"

We'll get into a practical example:
You have a limited budget and several advertising channels to choose from: social media ads, search engine marketing, etc..
Each channel has its own potential for success, but you're not sure which one will yield the highest return on investment (ROI).

To calculate the best allocation of exploring versus exploiting

1. Initial Exploration: Start by allocating a small portion of your budget to each advertising channel to gather initial data on their performance. Let's say you allocate 5% of your budget to each channel initially.

2. Collect Data: Monitor the performance of each channel over a set period, such as a month.

3. Calculate the math thingy

4. Prioritise Channels: The channel with the highest math thingy indicates the most promising opportunity for future investment.

5. Reallocate Budget: Allocate more budget to the channels with higher math thingy values (exploitation) while maintaining some budget for exploring new channels.

6. Iterate: Rince and repeat

But if you have a problem, where you can't "take back" your decisions, and can't reallocate resources, you're better off looking for the "perfect time to stop" looking for samples

1-2) Rule of 37%

I already talked about it in previous articles, but I really love it. But we'll give another example, the classic "hiring someone"

Having a correct sample size is critical - Source

Let's consider an example of using the "37% Rule" in the context of hiring for a position where you're interviewing 10 candidates.

1. Exploration Phase (37%): In the exploration phase, you'll interview approximately the first 37% of the candidates without making any hiring decisions. In this case, that would be the first 4 candidates. During this phase, your goal is to gather information about the pool of candidates and establish a baseline understanding of their qualifications, skills, and fit for the role.

2. Evaluation Phase: After interviewing the first 4 candidates, you'll have a better sense of what qualities and qualifications are available in the candidate pool. You'll use this information to evaluate each subsequent candidate relative to the ones you've already interviewed.

3. Exploitation Phase (Commitment): Once you've completed the exploration phase (interviewed the first 4 candidates), you'll use the remaining 63% of your interviewing time to compare each new candidate to the best one you've seen so far. If you encounter a candidate who is better than any you've interviewed before, you'll make a hiring decision and stop the interviewing process.

2- Under-commitement

Under-commitment is like being stuck in a perpetual loop of decision-making limbo. It's when you can't seem to fully commit to a decision, always second-guessing yourself or leaving the door open for constant revisions. This indecisiveness can be paralysing, leading to wasted time, resources, and opportunities.

Take my friend's case with the startup app. He started with technology A but then jumped ship to B when faced with a challenge. And who's to say he won't encounter similar hurdles with technology B? It's a cycle of indecision that can derail progress and drain energy.

AND OF COUUUUURSE IT'S ALSO a top 10 CIA tips to sabotage an organisation

And what's worse, is that trying to solve this, if done wrong, will make people stop raising issues, they will stop giving feedback and challenging decisions.

Which is really bad as it leads to confirmation biases and stagnation.
If you over-commit to a decision, you become stubborn.
Like Blockbuster over-committed to lending movies in the land of streaming services like Netflix leading to its downfall.

Tools to solve

Actually, it's no longer a data problem, it's a human problem.
The best thing you can do is:

2-1) Be flexible and open-minded

• Accept that you can never make the BEST decision and that's okay, try to do your best with it, and be open to being proven wrong.

• Embrace objections, people just want to help you make the best decisions, they are not "out there trying to get you". Most people are just trying to do their jobs, and their job is to be critical.

• But that doesn't mean every objection is valid, rank them by priority. If they can be dealt with after the decision is made, it's better (like the classic "people will not know how to use this", well you can start tutoring sessions AFTER the decision is made )

• Prepare an "opt-out" plan if possible, it's okay if something doesn't go according to plan, so just be ready if something important happens.

  • If it's a purely technological decision, it can also help you make your code depend less on the technology, making your code more modular which is great.

2-2) Make your decision process clear

• WRITE. IT. DOWN. Write the reasons, the process, the objections, etc… EVERYTHING. People will ask you about it in 6 months, you’ll be happy to have it around

• Explain to each person that will be impacted by the decision, why you made the decision

  • If you don't, people will always try to understand why are you making a decision, and will second guess your decisions.

• Decide on a time frame where you will 100% commit to a decision (major issues aside of course), and when to renew it.

• Make a Proof of Concept, It's great to showcase something without committing too much time to it (but don't use it as a real product)

• Explain to stakeholders that it's better to make a decision and course-correct if necessary than to linger in indecision indefinitely.

Closing thoughts

The past is an exact science. With hindsight, everything seems obvious. Looking back, it's easy to see where we went wrong or what we could have done differently.

Alright, wrapping things up here, let's remember that decision-making is like navigating a maze blindfolded – you're bound to bump into a few walls along the way. As we reflect on our past choices, it's easy to see where we took a wrong turn or missed a shortcut. The past is an exact science

Instead of dwelling on what could've been, let's focus on what we can do moving forward. Embrace the lessons learned, shake off the dust, and keep on memeing.

Life's all about trial and error, after all – we're just figuring it out as we go.