While immersed in a single initiative, one does not perceive the issues arising from the work. The data is prepared, definitions are set, and the team closest to the effort understands the context well enough to move forward. Problems that exist in the broader environment are absorbed quietly, handled through workarounds that become part of the process. No one flags them because they do not appear to slow anything down. They are simply part of how the work gets done in that specific setting.
When one has involvement in other initiatives, patterns begin to emerge. A digital access program depends on patient identifiers behaving consistently across facilities. A care management effort requires longitudinal views of patients built on definitions that shift depending on where the encounter took place. A financial initiative needs alignment between clinical documentation and how that activity is coded. Each of these programs is treated as its own workstream. Each encounters a version of the same friction, and each team resolves it independently using whatever method gets them to the next milestone.
Each team processes what they need to in order to accomplish their goal. It is not unusual. The challenge is to find the content in the work that tells you these teams are solving the same underlying problem in parallel, using different rules, often without knowing the other effort exists.
The problem does not present itself there. It presents itself later when the required outputs must be integrated across the portfolio. Timelines begin to stretch. Resource allocation shifts toward reconciliation rather than delivery. Project plans absorb weeks of effort dedicated to validating what the data actually represents before anyone can begin using it. Those weeks are rarely documented as a structural cost. They are folded into project timelines and described as complexity, when in reality they represent the accumulated weight of unresolved foundational issues being carried forward from one initiative to the next.
That takes time, and this is time that no one budgeted for or planned to do. It does not appear on any dashboard. It does not show up in a status report. It lives in the gap between what was projected and what was delivered, and over time it becomes the expected cost of doing business rather than being recognized for what it actually is.
In some cases, I have seen two separate teams work toward the same end goal and build entirely different data pipelines to get there, not because they chose to, but because neither team had visibility into what the other was doing at the data layer. The duplication was only discovered after both efforts were already in production. By then, the cost of reconciling them exceeded the cost of simply maintaining both.
The first thing to keep in mind is that this type of problem is not, and will not become, a visible problem unless you look at the portfolio as a whole. Individual initiatives continue to deliver. Teams continue to move forward. The friction is distributed across enough workstreams that no single one carries enough of it to trigger a red flag. It is only when you step back and ask how much collective effort is going toward resolving the same foundational conditions, over and over, that the scale of the issue becomes clear.
What you are working with is not progress in the way it appears at the surface. It is progress that carries an increasing amount of overhead with each new initiative, because the environment those initiatives depend on has never been addressed as a shared responsibility. Every new program inherits the same unresolved conditions and adds its own layer of workaround on top. Over time, the cumulative weight of that inherited work begins to limit what the organization can actually execute, even when the strategy, funding, and intent are all in place.
Once again you see the example where something that works well in a particular part of the system does not transfer cleanly into another. A model, a dashboard, a reporting framework, whatever was built and validated in one context begins to behave differently when it is introduced into a setting where the underlying data was produced under different assumptions. The output may still look reasonable, but the confidence behind it starts to erode, and the effort required to sustain that confidence grows with each additional environment the capability is expected to support.
The same problem is becoming more visible as organizations deploy AI at scale. A model trained on data from one facility encounters variation it was never designed to accommodate when applied across a system that has grown through acquisition or reorganization. The issue is not the model. The issue is the data environment the model depends on, and that environment reflects years of independent operational decisions that were never reconciled. AI does not create this problem. It inherits it and makes it harder to ignore, because the consequences of inconsistency show up faster and with greater impact when the decisions being informed are clinical or operational.
Addressing this issue is particularly difficult because it does not present itself in a way that is easy to measure. There is no single line item that captures the cost of unresolved data fragmentation across a portfolio. It shows up in delayed timelines, in duplicated effort, in the growing gap between what the organization plans to do and what it can actually deliver. Leaders who recognize this pattern tend to describe it in terms of friction or drag, something they can feel but cannot easily quantify or attribute to a root cause.
With time, this so-called normal work begins to shape what is actually possible to execute within the organization. While the strategy and initiatives continue to move forward, the gap widens between what appears achievable and what can actually be delivered. That gap is not the result of a single issue that can be corrected. It is the result of multiple recurring conditions, each carrying forward work that was never fully resolved and now requires additional effort to address.
Organizational efforts ultimately create friction by increasing the amount of time and energy required to move everything forward, without clearly exposing where the work is being held up. That is where the organization is most impacted, not in what is visible, but in what accumulates beneath it.
Related from Hutchins Data Strategy Consultants: Healthcare Data Strategy. On The Signal Room podcast: Rethinking Healthcare Data Strategy.