Booster Analytics

As a CTO, I’ve seen too many organizations in Asia, Australia, and Uk fall into the trap of believing that data is some magical oracle. Just collect enough of it, and answers will appear like a genie from a lamp. Spoiler alert: That’s not how it works. Data is not a crystal ball; it’s a resource. It doesn’t tell you what to ask — it just sits there until you know what you’re trying to solve. As an educator, I truly tried to imprint to my students to start with the problem.

Start with the Problem, Not the Data

Early in my career, I witnessed a company spend millions on a shiny data platform only to end up drowning in reports that no one could use. Why? They never defined the problems they were trying to solve. They believed the data would “speak” to them if they gathered enough of it. It didn’t. They had dashboards with hundreds of KPIs — but not a single actionable insight.

The lesson: Define the problem first. Are you trying to reduce customer churn? Optimize supply chain logistics? Predict equipment failures? Each of these problems dictates which data you need, how you should collect it, and what analysis framework will deliver actionable insights.

Case in Point: A Major Thai Bank’s Branch Dilemma

Consider a real-world example from Thailand’s banking sector. A large bank was grappling with a network of thousands of physical branches in an era when digital banking was rapidly gaining traction. The leadership team faced a major strategic question:

“Should we close branches to cut costs or repurpose them to support our virtual banking strategy?”

Many banks might have defaulted to collecting all customer interaction data and letting data scientists “find something useful.” But this bank approached the issue with a problem-driven mindset:

1. Defining the Problem: They framed the problem clearly — physical branches were costly, but closing them might hurt customer relationships in rural areas where digital adoption was slower.

2. Asking the Right Questions:

• Customer Data: Where are customer transactions still branch-dependent?

• Digital Adoption Data: What percentage of customers in each region are active on the bank’s mobile app?

• Service Use Patterns: Which services are most frequently requested in branches that could be digitized?

3. Data-Driven Decisions:

• Close or Consolidate: In urban areas with high digital adoption, they closed low-traffic branches.

• Repurpose: In rural areas, they transformed branches into “Financial Experience Centers,” providing personalized financial advisory services, digital literacy workshops, and remote banking support.

By focusing on the problem — balancing cost reduction with customer access — they avoided a blanket “cut-the-branches” decision. The bank ultimately cut costs while boosting customer engagement, showing how data works best when paired with a well-defined question.

Data Without Context Is Just Noise

Data by itself has no inherent meaning. The context comes from understanding the business problem. Take customer behavior, for example. A spike in website traffic might seem like good news, but without knowing your business context, that’s just a number. Is traffic up because of a recent marketing campaign or because a competitor’s website crashed, pushing customers temporarily your way?

I’ve seen companies waste months analyzing irrelevant metrics simply because “the data was there.” However, good data analysis is less about quantity and more about asking the right questions. Only then can you separate signal from noise.

Avoid the Data Rabbit Hole

Data exploration is a double-edged sword. It’s tempting to throw your best data scientists at the data lake and say, “Find something interesting.” However, without a clear business problem, this is just academic research disguised as innovation.

As CTO, I’ve made it a rule that every data analysis project must have a clear problem statement, hypothesis, and desired outcome. This keeps teams focused and prevents “analysis paralysis” — that endless cycle of finding insights that don’t matter.

Build a Culture of Problem-Driven Data Thinking

To succeed, organizations must cultivate a mindset where problems lead and data follows. This mindset requires a shift in how teams operate:

1. Business Leaders Define Problems: It’s their job to articulate specific challenges and business goals.

2. Data Teams Shape Solutions: They determine what data is needed, how to collect it, and how to apply analytics.

3. Continuous Feedback Loop: Insights must be tied back to the original problem, refined, and acted upon.

Final Thought: Data Is Not a Strategy — Problem Solving Is

As a CTO, I’ve learned that data isn’t valuable until you have a problem worth solving. Investing in data infrastructure before defining your business challenges is like buying the world’s best fishing gear without knowing where the fish are.

Stop expecting data to tell you what to look for. Define the problem first — and let the data follow your lead. That’s how real innovation happens.