The Comfortable Lie About Data Transformation
When I stepped into my role at Marshall University, we didn’t have a data problem.
We had dashboards.
We had reports.
We had talented analysts.
We had more data than most organizations know what to do with.
And yet, decision-making was still slow.
Leadership conversations often began with “Can you pull a report?” instead of “What should we do?” Meetings revolved around reconciling numbers rather than acting on them. Different groups brought different versions of the truth into the same room, and we spent more time debating who was right rather than debating strategy.
If I’m being honest, our first instinct was probably the same one many organizations have. We thought we needed better tools, ie, Better dashboards, more advanced analytics, new platforms, smarter technology.
That instinct makes sense. It’s logical. It’s comfortable. It’s also usually wrong if it’s your sole thought.
Because here is what I learned the hard way: if you simply bolt new technology onto old thinking, you don’t transform anything. You just get faster at doing the wrong things. This isn’t a story about dashboards, nor is it really a story about AI either. It’s a story about what happens when an organization decides to reinvent how it thinks, before it reinvents what it uses.
Adoption Without Advantage
Marshall’s data environment wasn’t a blank slate. We had invested in data over time. We had a home-grown data warehouse, established reporting processes, business intelligence tools, and people who cared deeply about accuracy and integrity.
But data was still viewed primarily as a reporting function, not as institutional infrastructure. Analytics lived in silos. Institutional Research did its work. Finance did its work. Enrollment did its work. Everyone was capable. Everyone was busy. And everyone was operating just slightly out of sync with everyone else.
Most importantly, data tended to show up after decisions were made, not while they were being shaped.
Leadership questions sounded familiar:
- What happened last term?
- Why does this number look different than that one?
- Can we get a report on this by next week?
Those are reasonable questions and necessary at times. But they’re also backward-looking questions in many cases. And when data is primarily backward-looking, it becomes a justification tool rather than a decision tool. We weren’t struggling because we lacked technology. We were struggling because we hadn’t framed the problem correctly.
Framing, Not Tools
There’s a well-known story about Kodak that’s often told as a cautionary tale about missing out on digital photography. But Kodak didn’t miss digital photography. They invented it. They adopted it. They sold digital cameras at scale.
What they missed was where the value was moving.
They treated digital photography as a substitute for film instead of a signal that the entire value system had changed. By aiming the right technology at the wrong objective, they accelerated their own decline. Most organizations don’t fail by ignoring technology. They fail by using powerful technology in service of outdated assumptions.
At Marshall, we were doing something similar, just in a much less dramatic way. We were producing more data, faster, with better tools, but we were still organizing ourselves around old questions, old workflows, and old power structures and not using analytics.
So instead of asking what new tools we needed, we forced ourselves to ask a much harder question.
One Question That Changed Everything
The real shift began when we changed a single question. We stopped asking, “What data do you want?” And we started asking, “What decision are you trying to make, and why is it hard right now?”
That sounds like a small change. It wasn’t. That question exposed issues we hadn’t been willing to confront:
- Decisions without clear ownership
- Decisions built on inconsistent definitions
- Decisions that relied on data arriving too late to matter
- Decisions shaped more by anecdote than evidence
- Decisions based on data without narrative or context
Once decisions became the focal point, everything else had to reorganize around them.
Data models changed. Timelines changed. Expectations changed. And, most importantly, roles changed.
From Reports to Decisions
The first major reframe was moving from reports to decisions. Instead of asking the names of our students and simply how many students we had, we started creating dashboards and asking:
- Which decisions mattered most?
- When were those decisions made?
- What information was needed at that moment?
- WHO our students were and HOW do we help them be successful?
This shifted our priorities quickly. Perfect data delivered too late became far less valuable than good data delivered at the right time. Relevance began to matter more than volume. Clarity mattered more than comprehensiveness. Data stopped being something that simply explained the past and started becoming something that shaped the future.
From Tools to Architecture
The second reframe was moving from tools to architecture. For a long time, we treated tools as the solution. But tools don’t create trust. Architecture does.
Trust came from:
- Shared definitions
- Transparent logic
- Reliable pipelines
- Clear governance
- And importantly, a unified data lake for dashboards
Governance was uncomfortable at first, and quite honestly still is. It sounded like bureaucracy. It felt like control. But in practice, it gave us speed.
When people trust the numbers, meetings move faster. When definitions are clear, debates become strategic instead of semantic. When data flows reliably, leaders stop hoarding their own versions of the truth.
Only after that foundation was in place did tools actually start to matter.
From Central Control to Distributed Capability
The third reframe was perhaps the hardest. Institutional Research was a bottleneck. Reports were reports but insights were lacking. So we built new dashboards with more capability. We built analytics that answered questions that led directly to efficiencies. Those dashboards have shaped our data and led to rapid change. We built capability.
Our goal shifted from answering every question to helping leaders ask better ones, from controlling analytics to enabling decision-making across the organization.
This redistributed power. And redistribution of power is always the real outcome of technological change, whether leaders acknowledge it or not.
What Reinvention Looked Like: Marshall for All
One of the clearest examples of this reinvention was Marshall for All. Marshall for All was not a dashboard project. It was a redesigned decision system on how to help our students succeed. The core question wasn’t “Can we track student outcomes?” It was “How do we intervene earlier, more consistently, and more equitably?”
That has forced us to rethink:
- What data mattered
- When it needed to surface
- Who needed to act
- How resources followed student need
Data moved upstream into action. It stopped being descriptive and started being operational, timelier, and has created true advantage for not just the university, but our students.
From History to Scenarios: Enrollment Forecasting
Enrollment forecasting was another turning point. Historically, forecasting had been about extrapolating trends. It was largely backward-looking. Reinvention meant shifting to scenario modeling. Instead of asking “What happened last year?”, leaders began asking:
- What happens if admits increase but yield drops?
- What happens if retention improves by two points?
- What happens if program mix shifts?
Data didn’t just inform decisions, it framed them.
What We Had to Let Go Of
Reinvention required letting go. We had to let go of hero analytics, or the belief that individual brilliance could overcome structural problems. We had to let go of one-off custom reporting that satisfied individual requests but undermined institutional coherence. And we had to let go of the belief that culture would change after tools arrived.
Culture had to change first.
Where AI Fits and Where It Doesn’t
AI is powerful. There’s no question about that. But AI doesn’t fix misaligned decision-making. It amplifies intent. If your intent is efficiency, AI will make you more efficient at doing what you already do. If your intent is reinvention, AI becomes a lever for creating new value.
The order matters.
Reinvention Is a Leadership Act
Technology does not transform organizations. Leaders do. Reinvention is not a data project. It isn’t a technology initiative. It’s a leadership act. Data simply reveals whether an organization is serious about it.
Here’s to being an effective and efficient storyteller, with data.
Brian M. Morgan
Chief Data Officer, Marshall University