Data-ing Is All About Customer Experience: Secrets Of Haunted Airplane
Budget-Friendly Ways To Build AI #5
I’ll be describing some core data-ing skills in practice and giving examples of how data-ing can really impact customer experience.
Data-ing Is About Being HANDS ON
With data-ing, one must spend time digging deep into the data. You have to get REALLY hands on.
The goals of data-ing are simple:
To learn the context
To dig into what anomalies mean and where they come from
To challenge or confirm your assumptions about the data
To find how the above 3 connect
To understand how to think about the above tips, let’s look at an example with service uses on flights.
A Haunted Airplane?
Imagine you’re looking at data on how airplane passengers use various services during flights. And you happen to notice that there are sometimes touchscreen usage on seats that aren’t supposed to be occupied. I’m very imaginative so I would immediately jump to the conclusion that the flight might be haunted. I may as well as start theorizing how the airplane’s electronic systems are wired impacts the way a ghost can interfere with it to cause anomalies.
If you are more realistic than I am, you might probably think it’s just a glitch.
But is it?
In reality, the true identity of this “ghost” is very likely someone moving around seats. If the plane is not fully occupied, people do move seats to find more room. In fact, I’m the kind of person who does this all the time! Even if the plane is very occupied, if there are empty seats you might still move around a bit to sit next to someone you know or to avoid a loud snorer.
It should be obvious here that a very important context to have is whether the flight is fully booked or not. If you want to dig deeper into the source of anomalies, you may as well look up which seats are occupied and how many empty seats are between them. You don’t need to have data on exactly why people move around seats but you do need to know some common reasons. If you’ve flown around enough, this may be common sense.
The 6th Sense of Data-ing
Talking about flights, here’s a story. On September of 2024, I was flying from the US to Seoul, and the flight happened to be about 60% occupied. This didn’t seem all that surprising as most flights between the US and Seoul are mostly occupied by Korean students who study in the US. Of course that means that flying to the US are occupied but flights from the US to Seoul are not.
I used to fly back and forth a lot between the US and Seoul as a student. When there are lot of rowdy teenagers and college-aged students, flight attendants pay special attention to crowd control. In practice this just means they are stricter about flight rules and etiquettes - not getting up when the seatbelt sign is on, asking passengers to be quieter if they’re being too loud, etc…
Since this flight was mostly calm adults - professionals, Korean-Americans visiting family in Korea during off-seasons, American tourists bringing an empty suitcase so they can load them up with K-beauty products, etc… - I presumed that the flight attendants would be pretty relaxed.
This was a late night flight and I happened to drink a lot of tea before the flight to stay awake and do some work. I was feeling a bit jittery so as soon as the plane took off and the seatbelt sign turned off, I decided to get up and use the restroom. I must have been in the restroom for like a good 20 seconds or so. And all respectful and reasonably hygienic people should know it takes at least 20 seconds to wash your hands properly.
This is what I thought was happening to me…
All of a sudden, someone starts slamming on my door asking “Sir are you ok? You need to please get back to your seat”. I asked why as there was no turbulence and the restroom had plenty of supplies. It turned out that it was a flight attendant. I complained to her about the situation but there was a very tall male flight attendant next to her asking me to get back to my seat in a firm way. Later they got me some hand sanitizers but this was really odd.
When I landed in Seoul and saw my parents, who were there to pick me up, I told this story to them. To my surprise, they looked at me like I was living under a rock and told me - “There was massive turbulence on a flight to Mongolia. Everything fell out and some people were injured. Of course they should be paranoid.”
The 6th Sense of Data-ing Is About Spotting The Hard To Spot Patterns
Data-ing requires a 6th sense because there can be one-off situations that can really do damage to customer experience or patterns of things that may not be so easy to spot.
Using The 6th Sense In Practice
If you are an airline, you must ask yourselves “OK How do we prevent customer experience from degrading?“. Where the 6th sense of data-ing comes in are like the following
Analytics Before Something Happens: If the airline knew what was going on with the weather situation per flight and communicated to the crew, this might not have happened. Or if someone in the cabin crew knew how the customers might react, they might have been able to handle things a lot more smoother.
Analytic After Something Happens: If there was an instance of customer experience dropping significantly, one has to look at how wide spread it was and why that was the case.
6th Sense = Contextualizing Your AI & Contextualizing AI Users: AI only knows what it knows. So it be fed contextual data and in some way. Whether that’s through RAGs to make AI understand context or feeding someone else’s data through protocols like A2A and MCP. All AI builders need to consider how their AI and AI users can develop this 6th senses together.
Conclusion: Data-ing and Dating Are Similar Cuz You Need to Spend Quality Time Together
Disclaimer: Being good at data-ing is not an indicator of being good at dating. But they both require you to spend quality time with what (or whom) is significant.
There is a delicate balance between being realistic and being imaginative about various aspects in your data. To think about how to approach this and construct your team, I would recommend reading my 2nd article about Great Data Engineering teams - there are lot of parallels.
If you are an AI Engineer, do remember that the most important thing to do is to spend quality time with your data. This means 2 things - 1) really getting to the bottom of topics I discussed in Part 1 of this article about avoiding various red flags that create cost traps later on ; and 2) Experiencing yourself what impacts the source of the data itself, in the case of airplane service use data actually being on an airplane.