Every company says data is important. It’s in the annual report, the strategy deck, the CEO’s keynote at the all-hands. “We’re becoming a data-driven organisation.” The phrase has become so common it’s lost all meaning.

Then you look at what actually happens. Customer records are scattered across five systems that don’t talk to each other. Marketing can’t tell you which campaigns drove revenue because attribution data lives in a spreadsheet someone updates monthly. The executive dashboard shows numbers that nobody trusts because the underlying data has known quality issues that never get prioritised.

This gap between what companies say about data and what they actually do with it is one of the most expensive problems in modern business.

The Collection Obsession

The first thing most companies get wrong is thinking the problem is not having enough data. So they collect more. Every click, every interaction, every sensor reading — all captured, stored, and largely ignored.

According to IDC, the global datasphere is expected to reach 175 zettabytes by 2025. Most organisations are sitting on more data than they could analyse in a lifetime. The bottleneck isn’t volume. It’s the ability to turn raw data into something useful.

A retail chain tracking 200 customer metrics but unable to answer “which products should we reorder this week” has a data problem, just not the kind they think. They don’t need more data points. They need better questions and cleaner answers.

Nobody Owns It

In most organisations, data responsibility is spread so thin it effectively belongs to nobody. IT manages the infrastructure. Finance owns the financial data. Marketing owns campaign data. Sales owns the CRM. Each department makes decisions about data quality, storage, and access independently.

The result is silos. Not just technical silos — cultural ones. Teams don’t share data because they don’t trust each other’s data. They build their own shadow systems to track the same information in slightly different ways. When discrepancies arise (and they always do), meetings get consumed by arguing about whose numbers are right rather than deciding what to do.

Some companies have tried to fix this by hiring a Chief Data Officer. It helps, but only if that role comes with genuine authority and budget. Too often, the CDO is a figurehead — responsible for data quality but without the power to change the systems or processes that cause quality problems.

The Dashboard Trap

Here’s a pattern that plays out constantly. A company invests in a business intelligence platform — Tableau, Power BI, Looker, whatever. They build beautiful dashboards. Executives love the visualisations. For about three months.

Then reality sets in. The dashboards show lagging indicators that everyone already knows. The data updates too slowly to be actionable. The metrics chosen for the dashboard don’t actually connect to the decisions people need to make. Gradually, attendance at the weekly dashboard review declines. Within a year, the platform is used by a handful of analysts and ignored by everyone else.

The mistake isn’t choosing the wrong tool. It’s building dashboards before understanding what decisions they’re supposed to inform. A dashboard should answer a specific question that leads to a specific action. “Revenue is down 8% this quarter” is information. “Revenue from enterprise accounts in Victoria dropped 12% because renewal rates fell after the service team was restructured” is insight that someone can act on.

Data Quality Is Boring but Essential

Nobody gets promoted for improving data quality. There’s no conference keynote about the time you spent six months cleaning up address records in the CRM. But bad data quality undermines everything else — every analysis, every model, every dashboard.

The most common data quality issues aren’t dramatic. They’re mundane. Inconsistent date formats. Duplicate records created by different systems. Fields left blank because they’re optional and nobody enforces them. Free-text entries where a dropdown should be.

Fixing these problems requires discipline, not technology. It means setting standards, enforcing them, and making data quality part of everyone’s job rather than delegating it to a team that runs quarterly “data cleansing” exercises.

What Good Looks Like

The companies that actually get value from their data share a few characteristics.

They start with questions, not data. Instead of “what can we learn from our data?” they ask “what decisions do we need to make better, and what information would help?”

They invest in boring infrastructure. Data pipelines, governance frameworks, quality standards. The unglamorous plumbing that makes everything else possible.

They treat data literacy as a core skill. Not just for analysts — for everyone. Product managers who can query a database. Sales reps who understand what a conversion rate actually measures. Executives who know the difference between correlation and causation.

And they accept imperfection. Perfect data doesn’t exist. The goal isn’t flawless datasets. It’s data that’s good enough to make better decisions than gut instinct alone — and clear documentation of its limitations so people know when to trust it and when to question it.

The Real Barrier

None of this is technically difficult. The tools exist. The methodologies are well-documented. The barrier is organisational. It’s the willingness to prioritise the slow, unglamorous work of building a genuine data foundation over the quick win of a new dashboard or the latest analytics buzzword.

Companies that clear that barrier gain a real competitive advantage. Not because they have more data than their competitors, but because they actually know what to do with it.