Why Attention To Detail Is An Important Trait

Many years ago when I lived in Las Vegas, I was walking inside the Miracle Mile Shops at the Planet Hollywood, Las Vegas and passed by a Ben Sherman store. Towards the back of the store, nearby the sales counter, there was a quote display on the wall that said:

Effortless cool is a detailed process.

-Ben Sherman

That quote stuck with me to this day. It’s a powerful quote that describes how most things that appear easy and effortless actually took a lot of work and detail to get to that level.

Most people don’t know that attention to detail is one of the things that separates excellence from mediocrity.

People become mediocre for one reason only: they’re lazy. And lazy people don’t pay attention to detail.

Data Scientists, AI Engineers, and data-driven business professionals often run into the problem of trying to present data to audiences in order to persuade them about some story or insight, but then their presentation falls flat. This leads to frustration especially when the one armed with the data has the knowledge about the truth, but the audience still isn’t convinced.

What’s even more frustrating is when one is armed with this data and knowledge about the truth, but this gives them no power or leverage in the company, and then the company lays them off. In the 2016 movie, “Batman v Superman: Dawn of Justice“, Lex Luthor (played by Jesse Eisenberg) masterfully said: “The bittersweet pain among men is having knowledge with no power…because that is paradoxical.”

In the end, CEOs eventually question their investments in data science and AI and will layoff these types of roles if they don’t feel they are getting a return on their investment. 

So how can you turn this tide and get leverage and power to prevent being laid off and increase job security as a data professional? After all, creating simple, effective data stories, visualizations, and products—seems deceptively easy. However, this apparent simplicity often leads to underestimating the complexity involved in achieving it.

Behind every elegant data visualization or intuitive data product lies a hidden, painstaking process. Crafting a compelling data story requires more than just compiling numbers, copying and pasting output from Jupyter Choke-books, and slapping it on some unformatted PowerPoint slides. It involves weaving a narrative that resonates with the audience and having an understanding of visual perception and meticulous attention to detail, from color choices to layout. 

In the realm of data science, the teaching of “Effortless cool is a detail process” refers to seemingly simple yet effective data storytelling, data visualization, and/or data product building which underneath is a deep, detailed process. This article goes into some of what that process entails as I share my experiences in my decade and half experience in data science and AI and managing large teams of Data Scientists, Data Analysts, and AI Engineers.

Data Storytelling: More Than Numbers and Raw Output

Creating an engaging data story involves more than just presenting facts, numbers, and raw output (especially if it’s not been formatted or straight out of a Jupyter Choke-book). It requires a strategic selection of data, understanding the audience and knowing what their pain points are, and weaving a narrative that makes complex information relatable and easy-to-consume. It’s also about finding the balance between too much information and just enough to tell a compelling story. This process is far from effortless.

There are a few Axioms from Remix Institute’s Axiomatic System of Philosophies that can help guide you in crafting a better data story. These Axioms are things I’ve learned in my decade and a half experience presenting to senior executive and C-Suite audiences that have allowed me to succeed in influencing and persuading them.

1 – Occam’s Razor and Principle of Parsimony. Simplest explaination is the best explanation possible. Less is more. Make your data story easy to consume and understand. This concept was first introduced in the 14th Century by William of Ockham, a Franciscan friar and theologican, and it has withstood the test of time. 

2 – Entertain Your Audience. Let’s be honest, most people are just looking to be entertained, including senior executives and C-Suite professionals. It’s why sports, movies, and social media are the dominant content that gets consumed. People just want entertainment, and they’ll always just want entertainment. The sport of gladiator fighting ran from about 105 BC to 404 AD, which is proof of its popularity in the Roman Empire’s entertainment calendar. The concept that people want to be entertained is also another Axiom that has withstood the test of time. Remember when you were a kid and you were entertained from stories in kindergarten pop-up books? That’s what you need to do with your data stories or you will put your audience to sleep and lose interest.

3 – ELI5 Model or Explain It Like I’m 5 Years Old. Distill complex information into easy-to-understand concepts. Do not assume your audience has any prior knowledge about what you’re talking about especially if it pertains to artificial intelligence, machine learning, statistics, mathematics, databases, etc.

4 –  Show Them Only What’s Necessary. Presentations get derailed for one reason. If you memorize this reason, I guarantee you that your presentations will have a much higher success rate. Here is the reason: Anchoring leads to the Law of Triviality which subsequently leads to Analysis Paralysis and then no decisions get made. Put simply, do not tell the audience to think about an elephant because then they will think about an elephant. And then they will anchor on the elephant and give disproportionate weight to the elephant even though the elephant is trivial. And once they’ve given disproportionate importance to the trivial elephant, then they’ll send you down rabbit holes and have you complete analysis after analysis until they finally reach paralysis. Then no decision gets made, and progress gets stalled.

Data Visualization: The Detailed Art of Clarity, Simplicity, and Minimalism

An effective data visualization might look simple, even obvious after observing it. However, the path to this simplicity is anything but. It involves a deep understanding of visual perception, color theory, and the art of highlighting what’s important while eliminating noise. Each element in a visualization – from font and color choices to the scale of axes – is a deliberate decision aimed at making the data accessible and understandable. This process demands a keen eye for detail. 

In the 2020 book “Shikake: The Japanese Art of Shaping Behavior Through Design” by Naohiro Matsumura, the author teaches about shikake (or “device” in Japanese) which is a design that exerts influence on people through subtlety, rather than a direct prompt. By combining traditional and minimalist Japanese aesthetics with lessons from behavioral economics, shikake is designed to encourage a certain behavior without telling the (often unwitting) person the primary purpose behind that behavior.

IKEA stores and Swedish architecture are also able to subtly nudge customers to perform certain behaviors. This level of detail is far from easy.

Let’s give an example. Let’s say I wanted to present a simple visualization to my audience to get them to look at my chart. I could prompt them to say, “take a look at this chart and analyze these numbers.” Or I could spend just a little more time sprucing up the design of the chart while still keeping it minimal in order to make it easy for them and subtly nudge them where I want them to go. The latter will get the audience to stare at the visualization longer, forcing them to actually look at the chart, versus just telling them to look at it. Audiences want to be entertained: they will not look if they get bored.

It’ll also prevent then from just glancing at your chart and not grasping whatever story you wanted to tell. The key is to guide their eyes where you want them to go and have them stare at your art longer.

Most Data Scientists and Data Analysts will build something like the chart below because it’s the easy and lazy to do. I’ve seen this type of chart countless times when I had to review presentations from employees I managed before meeting with senior or C-Suite executives. It’s also common to see on tutorials and blogs so people tend to copy it:

But if you added just a few more details to the chart to make it easier to consume and add just a few more lines of code, you get something that will cause the audience to stare at it for longer:

What is the difference between the two charts?

The first chart is full of common mistakes I see being built by Data Scientists and Data Analysts who then try to present it to their audience:

• Not labeling the x and y axes in user friendly labels; they just use the raw value names. The improved second chart explicitly labels the axis in easy-to-read labels.
• Using the standard, out-of-the-box theme from ggplot2, matplotlib, et al without customization. The improved second chart uses themes and customization.
• Not changing colors and fonts to match brand colors. The improved second chart changes the color, fonts, and uses branding.
• Not using the chart title to tell the story so that the audience doesn’t have to think. The improved second chart just tells the story without making the audience think about what’s being displayed.

If you avoid these mistakes and just add a little attention to detail about how your audience will consume your data visualization, then you’ll become effortlessly cool and influence more decisions. 

Data Product Building: Intuitive Design and Ease-Of-Use

Building user-friendly data products is no less challenging, necessitating continuous refinement to ensure seamless functionality and design. This involves taking a deep dive into user needs and behaviors. The aim is to mask the complexity of data behind an intuitive and user-friendly interface.

The main thing I want to emphasize is regularly getting feedback from your users to enhance your data product. This means communicating frequently with your users. You can’t just set-it-and-forget-it. Continuous improvement of your data product based on your users’ feedback will increase its usage and adoption.

You literally have no choice but to communicate regularly with your users or your data product is dead in the water. When I managed teams of Data Scientists, Data Analysts, and AI Engineers, this was the key differentiator that set apart my top performing employees from my bottom performing employees. The top performers communicated regularly with their users and built for their users in mind. The bottom performers just wanted to build in isolation and not talk with their users.

It was also one of the key differentiators that set apart top data science teams I was a part of. The top data science teams had a direct line of access to their users and stakeholders, especially the C-Suite and senior executives. You have to get your work known and get your users involved in its creation. 

Here’s some tips I’ve learned about building great data products over the past decade and a half that have helped ensure the success of my data science team’s initiatives and elevate their reputation:

• Single page applications trump multiple-paged and multiple-tabbed applications every time.
• Pleasing aesthetics increase user adoption. Design for the user in mind.
• Do not duplicate information on your dashboard or data product.
• Minimize the number of clicks a user has to make on your dashboard or data product.
• People consume data products visually in the following direction: Top-to-bottom and left-to-right. 
• Do not use complex visuals. Don’t create work for your users.
• Don’t use Pie Charts. Ever. 
• Don’t ever have more than 4 time series lines on a Time Series Line Chart.
• If you’d like design feedback on a dashboard or app, consult a UI/UX expert at your company. Many times, they may even help you wireframe something that looks top notch.
• Collect feedback from at least 5 users of your dashboard or data product. Research has shown that 85% of your usability issues will be uncovered here. Links to that research can be found here and here.

Conclusion: Effortless Cool Will Make You Highly Valuable

Achieving “Effortless Cool” is indeed a detailed process, but if you master it, it will make you more valuable, more in-demand, and most importantly, more persuasive. Effortless Cool represents the mastery of making something complex more intuitive, engaging, and seemingly simple.

Aspiring data professionals should appreciate the meticulous work behind this façade of effortlessness, recognizing that true skill lies in the ability to make the complex appear beautifully simple. To truly master the art of ‘Effortless Cool’ in data science and AI, professionals must embrace the detailed process behind it which means recognizing the importance of crafting narratives in data storytelling, understanding the principles of design in data visualization, and focusing on user experience in data product development.

It’s not easy to do, but the payoff is huge.

Remix Institute Membership

If you liked this article, join our free Membership as a Stádas Genesis member for access to elite professionals and exclusive courses on getting started with R, Julia, data science, AI, and six-figure job opportunities. Elevate your skills and gain an edge in the industry. Sign up now for the elite learning experience.