Will you trust the data?

DATA
LIFE

All of us make decisions almost everyday. But what affects the decisions that we make? Is it the training that we received by the action-consequences loop? Is it the 'Mitochondria is the powerhouse of the cell'-like facts? Is it the anti-'The Road Not Taken' pressure from society? It's often a mishmash of all these together and more – different recipes for different decisions. When I decide to jaywalk to cross the road, it's mostly because that's the only practical option as a pedestrian. Since humans haven't evolved flight beyond a few feet of jump, and Indian roads haven't evolved pedestrian crossing beyond faded white stripes ending in the road median, we're left with only the overconfidence of the signalling of our palms stretched out to the drivers coming from the right side of the traffic.

This life-at-stake decision would've been unimaginable if I didn't have the data. I see many people crossing across the dangerous flow of tins on wheels. I've seen some close calls (been part of some), seen frustration of both parties (been part of some), seen the ad-hoc group forming to appear bigger to the enemy (been part of some). But I've never seen an 'accident' (calling it an accident is like deliberately running into someone after stalking their schedule, and calling it a coincidence). Hundred percent of the pedestrians who I've seen wanting to cross a road (even highways sometimes), do cross the road. That's a hundred percent success rate discounting the information from the news which will make it a few points lesser. It may take 5 minutes for the flow of traffic to suit your risk-appetite, but it's still an almost hundred percent success rate according to my brain.1

Okay, so I decide to cross the road based on human data. How about deciding to be or not to be? This one is easy (for me). To be is the answer, and not to be isn't even a close second choice. Not being relieves the daily grind, and the awful things that we sometimes go through, at the cost of cancelling all the fun parts of life too. But by the virtue of free will vested in me, I can choose a different grind, or more rejoice, or at least turn a few dials which can make the fun-to-not-fun ratio at least better than 0:0 (sorry math nerds). So, this decision is easy just by the belief in the existence of the tunable dials (who knows how hard some of them might be to turn though).

Another example is quitting your job. If I tell you that 100% of average candidates get a job with a higher pay once they've quit the job they hated, will you quit your current job that you've been nagging about? It may take a few month of job hunt, but it'll be better than your current pay. Will you trust the data and take that leap of faith (in data)?

What about going to the gym? There's undeniable data about exercise and diet regimen helping a person become healthier and more attractive when they follow a certain set of movements and eat a planned set of nutrients. But will you trust the data? Many people do and follow the path, only to start doubting the data when the results don't show. But the data is out there, barely changing, too cool to care about the opinion of the sheep.

It started as a thought experiment in my brain. There are times when we need data. For example, if you're starting a business, the data about market demand and trends is essential to assess the risk-reward ratio. When you're hiring a candidate, the data about their past accomplishments, and current skills is important. When you choose to spend seven lifetimes with someone, you need the data about their personality, and values. So, the data seems to be playing a key part in solving the decision dilemma. But will you trust the data?

The answer to the question isn't easy. We will trust the data under a few conditions.

First and most obvious condition – the data should be true. This doesn't need much explanation. False data will throw off the reasoning for the most part.2 With a sound reasoning, wrong data will cause wrong decision. And ambiguous data is as good as no data. Imagine, the weather forecast before you leave the house. The weatherman enthusiastically tells you to put extra sunscreen because the UV index will be up 2 points from yesterday. You follow the advice and leave the house all set to beat the UV rays confidently. But your confident doesn't match those actresses in the ads who don't need a dupatta covering their face because they've got 0.1mm of SPF 50 PA+++ cream applied. And this lack of confidence was not in your skin care, it was in that sly weatherman in his gray suit who told everybody in your area to protect against the sun; because it's raining now, and you can't do anything because you were fooled by false data.

Second condition: the data should be enough. This is less intuitive, and many people fall for this trick. And this is probably the most common way to lie without lying. Imagine, you are a woman swiping on one of the niche dating apps even though this one is also owned by Match Group and has nothing different apart from the UI. You're tired of flicking your thumb in just one direction. After countless reps of thumb curls, your see a profile that checks all the boxes. He is handsome, he is funny, he works in the creative industry and has actually got the hobbies that you thought could only be described as just the right mix of nerdiness and instagrammability for you. Your thumb does an awkward jerky swipe after all the muscle memory. But no mishaps so far. You stalk him on LinkedIn just to be sure (who am I kidding, you stalk him everywhere, ChatGPT's Deep Research isn't anywhere close to your skills). You text for a few days and find a cute little cafe for a date. He arrives in a car which you knew he owned. He also took the effort to reach on time from his home (which he owns) to this cafe that you like. His photos weren't lying. The data was amazing, and you've already imagine all the hunky-dory scenarios in your head. The data so far is true, but unfortunately wasn't enough. You pickup the newspaper the next day to see the same guy in the newspaper who has already been married for 5 years, and is now being charged of domestic abuse. This piece of missing information completely overrules all the previous data.

Third condition: the data should be reliable. This is closely related to the first condition that the data should be true, but it's not the same. Reliability of data is hard to ascertain. For first-hand experiences, we have the highest reliability. For example, burning your tongue on a hot coffee. You know how long it lasts, how painful is it, and how Starbucks isn't going to refund you for what's left of it and you have to just blow and sip the rest of it just to make your money's worth. On the other hand, if a stranger on the street tells you data, it is almost unreliable. The stranger could be an angel that wants you to become rich, or the stranger could be a snake oil salesman that wants you to think that you can become rich. I'll give you more concrete examples. Companies almost always throw data at their customers like this amazing toothpaste that can revert cavities according to their lab, or this great college whose students got big placement offers, or a social media platform which tells you how people are more likely to engage in their platform's ads instead of their competitors. The reliability of this data is hard to assess. And I believe that the lack of faith in data is what keeps most of us from trusting and acting on the data.

Once you have these pieces together, you will probably make better decisions. But I think the role of data in decision making is crucial, but it's also overestimated in many cases. Look at the dilemma of working for an employer, or starting a business of your own. Even with the data, you can't make the right decision. Why? Because both choices can be right, and both can be wrong. It doesn't have to be the exclusive or. Sometimes any choice can be right/wrong, and you can't do all at the same time. So, when you go down one path and it turns out wrong, you think the other path was right. No! Maybe the other was was even worse. Or the other way around when you go down one path and see it working out well. You might think the other choice was wrong, but maybe that was an even better choice. These confounding dynamics are caused by what's called not-being-able-to-predict-the-future. Some things in the world are predictable, but most of the complex decisions involve variables that are really hard to predict. There are "experts" who people like to turn to for predictions, but many experts are experts because people just call them so. Like the stock market "experts" who got beaten by a lazy index fund.3

I can't solve this dilemma for crucial decision, but I have a theory for low-to-moderate stakes decisions like going out with friends, buying a new gadget, or asking somebody out. I'm sure many of us think too much and do tool little. We dwell too much analyzing each piece of information and play out the two timelines we can think of. For such cases, we should trust the data keeping in mind some basic sanity, but we shouldn't become data scientists in our head for such decisions. We should take a leap of logical faith, and take it quick. If you don't make a decision, somebody/something will take that decision for you. There's a story of Ronald Reagan about indecision. In short, he had to get a pair of shoes made but couldn't decide if he wanted a round toe or a square toe, and he kept putting the decision off whenever the cobbler asked. So one day, the cobbler finally made his shoes – one shoe had a round toe and one had a square toe.4

This is the first long post that I've written. I used to write very short posts on LinkedIn because I know what our attention span has become, but writing short posts is harder and they're not fun to read either. I want to write more, and I also need feedback. Thank you for reading.

Footnotes

  1. We'll keep dogs and rats away from this discussion because there are cases which have shown me less than 100% crossing rates, and unfortunately similar survival rates. ↩

  2. If you're reasoning is also wrong, then maybe the effects will cancel out, making it a right decision but we'll talk about reasoning some other time. ↩

  3. Derek Muller's video on experts is a good one. Veritasium: The Expert Myth ↩

  4. The Cost of Indecision ↩