“Why ninety days?”

“Because that is the number we use.”

“But who actually signed off on ninety days?”

“Nobody signs off on gravity, Wei; it is just there.”

The air conditioning on the twenty-third floor near KL Sentral always has a slight metallic tang to it, like someone once left a bag of pennies in the vents. We were sitting around a table that felt too large for the four of us, staring at a slide that defined “Active Churn.” On the screen, a red line dipped toward a RM 500 threshold.

Someone from Marketing leaned forward, their shirt crisp and their face displaying a level of confidence usually reserved for people who know where the exits are in a burning building. They asked why we exclude accounts under RM 500 from the primary churn report. The Finance lead, a man who looked like he had been built out of old ledgers and stubbornness, didn’t even look up from his coffee.

He said the threshold came from the previous system. He explained that the old SQL server would time out if we tried to pull a report that included every single micro-account. That system was decommissioned in 2017. We are currently staring at a RM 500 filter that exists purely because a server in a basement five years ago didn’t have enough RAM to finish a task.

Ancient Fossils and Thursday Afternoons

Most of what we call “business intelligence” is actually the study of ancient fossils left behind by people who were just trying to get home on a Thursday afternoon. I bit my tongue earlier this morning while eating a piece of toast-a sharp, dull throb that makes every word I type feel like a small betrayal of my jaw-and it has left me in a state of heightened irritability regarding things that don’t make sense.

If my tongue hurts for no good reason, I at least want my data to have a pedigree. But it doesn’t. Your company’s most sacred KPI was likely invented by a temporary intern in 2013 named Wei Han who needed a “placeholder” value for a column so he could export a CSV before his bus arrived.

He chose ninety days because it felt like a round number. He chose RM 500 because it was the price of the phone he wanted to buy that weekend. Ten years later, a board of directors is making a RM 20 million pivot based on the ghost of Wei Han’s desire for a new smartphone.

A rusty stapler on a mahogany desk is often the only thing holding together the fragile fiction of corporate wisdom. We assume that because a number is in a slide deck, it has been vetted, peer-reviewed, and blessed by a council of elders. In reality, numbers acquire a strange kind of gravity.

Once a definition is written down, it starts to collect weight. People build reports on top of it. Then they build dashboards on top of those reports. Eventually, someone builds a bonus structure on top of those dashboards. At that point, the RM 500 threshold is no longer a technical limitation; it is a load-bearing wall.

To change it would be to admit that the last six years of “growth” were actually just fluctuations in how we filtered the small stuff. Nobody wants to be the person who tells the CEO that the “Record Year” of 2021 only happened because we forgot to update a definition. The past survives.

The Horse’s Backside of 2013

There is a famous industrial anecdote about the width of the US railroad tracks being exactly 4 feet, 8.5 inches. It seems like a random, oddly specific number until you trace it back through English tramways, which were built by the same people who built the pre-railroad wagons.

Those wagons used that specific wheel spacing so they wouldn’t break while traveling on old long-distance roads in England, which had ruts already worn into them. Those ruts were originally carved by Roman war chariots. The Roman chariots were designed to be just wide enough to accommodate the backsides of two war horses.

This means that the design of the Space Shuttle’s solid rocket boosters-which had to be shipped by rail through a tunnel that was only slightly wider than the tracks-was ultimately determined by the width of a horse’s rear end two thousand years ago. Your business is currently running on the horse’s backside of 2013.

In the Malaysian market, where digital transformation is often discussed with the same fervor as a new nasi lemak stall opening in PJ, we tend to focus on the “digital” and ignore the “transformation.” We buy the most expensive tools, but we feed them the same prehistoric definitions.

We use Power BI to visualize a lie. We create beautiful, shimmering dashboards that are essentially high-definition maps of a territory that doesn’t exist. This is the gap that separates a “user” from an “analyst.” A user drags a field into a bar chart and calls it a day. An analyst asks why the field is filtered to ninety days and refuses to move until they find the intern’s ghost.

The Act of Rebellion: Semantic Modeling

To fix this, you have to move away from “reporting” and toward “semantic modeling.” It sounds like academic jargon, but it is actually an act of rebellion. It is the process of taking those hidden, accidental rules and writing them down in a way that is visible, governable, and-most importantly-changeable.

When you use a tool like Power BI properly, you aren’t just making charts. You are building a layer of truth that sits between your messy, chaotic database and your decision-makers. You are defining what a “Customer” is in a way that doesn’t depend on a server timeout from 2017.

This is exactly what we focus on during power bi training at Trainocate Malaysia. It isn’t just about where to click or how to make a pie chart look pretty-though we do that too.

It’s about the PL-300 curriculum, which teaches you how to design and manage analytics solutions that actually hold up under questioning. It’s about learning DAX (Data Analysis Expressions) so you can write a measure that says, “This is Churn, and here is exactly why it is RM 500,” or better yet, “This is Churn, and we are changing it to RM 400 because the data actually supports it.” It is about moving from the archaeology of the past to the architecture of the future.

If you are an analyst in Kuala Lumpur or Penang, or if you are managing a team that is currently arguing over a RM 500 threshold in a KL Sentral boardroom, you have to realize that your inherited models are failing you. They are heavy. They are slow. They are full of “calculated columns” that someone added in 2019 “just for one specific report” and never deleted.

These models are like a house where the previous owner decided to bury a car in the backyard and then built a patio over it. Every time it rains, the ground sinks, and nobody knows why.

It proves you understand the star schema-the fundamental structure that makes data move fast and accurately. It proves you know how to handle row-level security so that the Finance lead sees one thing and the Sales team sees another, without creating twelve different versions of the same file. For Malaysian companies, this isn’t just a “nice to have” skill; it is a survival requirement.

With HRD Corp claimable courses, the barrier to getting a team trained is lower than it has ever been. The only real obstacle is the comfort of the old definition. I look at that red line on the screen again. The throb in my tongue has subsided into a dull ache, a reminder that mistakes have consequences.

From sacred relics to living documents

If I don’t pay attention to how I chew, I get hurt. If we don’t pay attention to how we define “Active,” the company gets hurt. We keep talking about “Data-Driven Decision Making,” but if the data is just a collection of historical accidents, we aren’t being driven by data.

The goal is to reach a point where, when someone asks “Why ninety days?”, the answer isn’t a shrug. The answer should be a link to a documented semantic model where the logic is laid bare. It should be a deliberate choice based on current market conditions, not a legacy of a system that was switched off before the current interns were out of secondary school. We have to stop treating our reports like sacred relics and start treating them like living documents.

Every report you see today is a story. Some of those stories are true, and some are just old rumors that nobody bothered to fact-check. The difference between a successful organization and one that is slowly sinking into the mud is the ability to tell the difference. You cannot build a modern empire on a horse’s backside. You cannot lead a market if your definition of “success” is limited by the memory of a server that died in 2017.

We ended the session at KL Sentral without changing the RM 500 threshold. It was too “risky” for the quarterly report, they said. But as we walked out toward the trains, I saw the Finance lead pull out his phone and look at the churn dashboard. He looked tired.

He looked like a man who knew he was looking at a map of a city that had been torn down and rebuilt three times since the map was printed. He deserves a better map. You deserve a better map. It is time to stop digging up fossils and start building something that actually stands on its own.

Change is expensive. Not changing is more expensive.

Explore power bi training at Trainocate Malaysia