The following story is based on real experiences. Names and details have been changed to protect confidentiality.
Imagine a healthcare service provider. Let’s call it “Health&.” Frontline staff there use a custom digital platform to input patient information: everything from medical history to service plans to follow-up appointments.
Meet Sarah, a dedicated data engineer at Health&. Each morning, Sarah opens her laptop and faces the same frustrating reality. There are hundreds of column headings in a database, many she doesn’t fully understand. It’s also unclear why the numbers can shift wildly between the columns.
“I’ve been staring at some of these field names for years,” Sarah shared during a meeting, adding something like, “I know ‘patient_status_4’ is important because people keep asking about it, but I have no idea what triggers a change in that field or how it’s different from ‘patient_status_3’.”
The stakes were high. Her confusion meant the data and subsequent decisions, logistical planning and patient care may be wrong and costly to the organization and its customers.
Sarah is responsible for preparing data that her teammates use to create dashboards for executives, managers and others to make critical decisions about internal operations and patient services. When those decision-makers spot unusual patterns or need to understand relationships between different data points, they call Sarah’s manager, Maya.
“I constantly get urgent requests to explain the data,” Maya shared. “Then I have to go to Sarah, who has to try and figure out what’s happening because it depends on how the frontline staff enter this information when they’re with patients.” Sarah seeks out answers, but the platform keeps changing. It sucked her time and energy.
The Invisible Disconnect
When Sarah couldn’t properly interpret data fields or why data behaved strangely, the resulting dashboards could present misleading information. Executives might make decisions based on misunderstood data, potentially affecting thousands of patients and millions in resources.
Maya remembered an incident like: “We thought patient completion rates had suddenly dropped by 30% in the Eastern region. Executives were about to reassign staff and restructure the whole program when we discovered it was just a change in how data was being categorized. They almost made a catastrophic decision based on a misunderstanding of the data.”
The root cause? The product team that built and maintained Health&’s digital platform focused on the needs of frontline staff. They never considered backend data workers like Sarah and Maya as users of their system.
“The platform works for data entry,” noted Rahim, a frontline staff member. “We had no idea the data was so confusing on the backend.”
A Rapid Researcher Stumbles In
This is where I entered the picture at Health& by doing UX research on their data dashboards. I conducted standard discovery interviews, but the importance of this disconnect between the UX of frontend and backend teams was not clear.
Then one day Sarah sent a message to the team chat, “I need to go to frontline staff school.”
I pinged her. She shared not knowing why there is a change among patients from “patient_status_3” to “patient_status_4” in the data. Explaining this to leadership was stressful. I was shocked at how much time she and her teammates spent trying to decipher data field meanings and the patient process as part of the work. The downstream effects of the product’s design were enormous: delayed insights, potentially flawed decision-making and countless hours of wasted work.
Sarah said something like, “We’re making educated guesses about what these fields mean and what patients do, but we’re never completely confident.”
Building Bridges Across Teams
It seemed our product team hadn’t connected the experiences of the frontline teams building the product and entering the data with the data engineering, management and analysis teams on the backend. I was only able to bring some of them together by arranging a meeting between Sarah and Rahim, who had entered patient data for years. Sarah and I prepared for this meeting:
- She identified all the confusing data fields
- We organized them in a spreadsheet according to her understanding of the process Rahim and patients followed using the product
- We asked Rahim to come prepared to show us the actual pages he entered data on during patient interactions in order to clarify the process for us
When we all met for an hour, Rahim walked us through each page of the application, explaining the patient journey and what triggered changes in each field. Sarah’s eyes widened as mysteries that had confused her for years were suddenly clarified.
“So status_3 actually means the patient has temporarily paused services. Status_4 means they unpaused services,” Sarah realized. “That explains why we see so many patients moving back and forth between status_3 and _4.”
Sarah and I then visualized how patients moved through the system and what caused data to jump between fields or column headings. This map transformed Sarah’s ability to interpret the data and explain it to leadership.
“This saves me hours every week,” Sarah later told me. “And the dashboards being created better reflect what’s happening with our patients.”
Why This Matters for Your Organization
If you work with data systems or digital products, you might recognize this situation. Data workers often struggle silently, trying to make sense of information from systems they weren’t ask to help design and don’t fully understand.
The costs of this disconnect can be substantial:
- Wasted hours spent deciphering confusing data
- Delayed insights that slow organizational decision-making
- Potential misinterpretations that lead to poor strategic and logistical choices
- Frustrated data teams who can’t fully apply their technical skills
This sort of story is not new. People have been building data products for years with similar challenges, but it can be a light lift to change course.
Proven Solutions
Start with connection. Bring together the people entering data and developing the product with those building the data systems, managing and analyzing it and any other operations folks. I was limited in my ability to connect them all, but the connections I made mattered.
Organizations can save countless hours of confusion and frustration by creating these bridges. Data quality can improve significantly as the people managing and pulling the data understand which fields are necessary to populate dashboards to answer questions for executives, managers and frontline teams.
Use rapid research and facilitation to:
- Identify the disconnects from data input teams to data management, analysis and visualization teams as well as the different job functions using the data to make decisions
- Surface feedback-sharing opportunities between these teams at already scheduled meetings
- Create visual documentation to clarify data fields and their relationship to each other other and frontend processes, products and platforms
- Develop sustainable practices for maintaining data quality as systems evolve through periodic strategy sessions and UX research on any product that data teams touch
As a seasoned practitioner who has done this work, I’d be happy to think through adapting these approaches to your context. Let’s chat.
Note: An AI assistant helped me write this. I checked its database and elsewhere for any plagiarism or misattribution. If you identified anything not cited correctly that should have been, then please let me know. Sharing this might sour your reading experience, but recognizing others’ ideas is also important. People are exploring how to live with AI responsibly. Shouldn’t we try figuring it out together?