Mistakes that Cause Most Data Projects to Fail
Beginning a significant data project can be overwhelming, especially starting from the ground up.
In most organizations, accessing data is time-consuming, confusing, and requires a map to navigate the myriad of levels and pathways to reach the information you seek.
In some cases, information may be stored in more than one system, even in more than one electronic health record system. Once accessed, data is often complex, static, and not user-friendly. Not only that, but when not properly combined, specific sources seem to have no connection or purpose.
You aren’t alone in this struggle. According to a study done in 2017 by Gartner, 60% of big data projects fail to go beyond piloting and experimentation and will be abandoned. As bad as that sounds, according to Gartner analyst Nick Heudecker, Gartner was "too conservative" with its 60% estimate. The real failure rate? "Closer to 85%."
What causes so many of these projects to fail? And how can your organization prevent making the same mistakes?
What Causes Big Data Projects to Fail?
Pinnacle Health Informatics has specialized in extracting and integrating behavioral health data for over two decades. During our time in the industry, we have seen our fair share of data projects. As we have assisted our clients, we saw six common mistakes that were causing these projects to fail.
Difficulty accessing data / Data in an unfriendly structure
Lack of in-agency skills
Lack of a Champion
Poor data quality
Difficulty accessing data/data in an unfriendly structure
Data stored natively in an electronic medical record must, by necessity, conform to strict rules of 'normalization.' Data will only live in a single place in a database. This system works well for computers but can be confusing and difficult for humans to understand and work with.
When working on a large data project against the electronic medical record production database, one can often spend more time hunting for the data they need than actually building meaningful reports and dashboards.
Spending so much time hunting for items can wear anyone working on a large data project down and leave them unproductive. Because the Pinnacle Data Warehouse structure is human-friendly, data analysts, report writers, and end users can spend more time getting valuable data and making sure decision-makers get the data they need.
This results in faster development, better access to data, and ultimately, better analytics and patient care.
A data project with unclear objectives is destined to fail. Clear objectives and goals are vital to a successful data project.
Do you want to increase productivity in your workplace? Do you want to monitor CCBHC data? Do you want to improve patient outcomes?
We like to start each large project or contract with an interview involving the stakeholders and discuss, in a non-technical way, what they’d like to accomplish with an eventual data visualization or process. We’ll ask how success will be measured and what exceptions and opportunities the tools will address.
Clear objectives and goals will set you up for success in your data project.
Poor planning can cause consequences for any large project —from missed deadlines and budget overages to team burnout and client frustration. That’s why it’s essential to establish a solid process you can use to plan your data project.
Before you start anything, bring your team together and create a detailed project plan that outlines everything from the data collection process to the final reporting and analysis to ensure everyone is on the same page.
Lack of in-agency skills
One challenge that many organizations run into with big data projects is a lack of skilled personnel. The lack of skills in organizations contributes to 30% of the failure. This can affect the organization at various levels.
Unfortunately, when an organization is yet to embrace data as a culture, many personnel see data as complex, overcomplicated, and not worth learning to navigate. These individuals are likely to squander promising business opportunities and often fail to see problems until these problems become full-blown crises.
The challenge lies in ensuring Big Data projects perform reliably and efficiently enough so that organizations can flip their mindsets from considering Big Data only as a defensive tool for current activities to using it as a catalyst for business growth.
This can be accomplished by embracing data as a culture of your organization that will continue to evolve and change with time as you learn more. If you develop a data culture, then data is ever-present. A data culture equips your organization with the insights they need to be truly data-driven, tackling your most complex business challenges.
Lack of a champion
A successful data project requires an individual inside an organization to take ownership of the process, maintain a keen interest in its success, and be a visible and vocal champion of the process.
Like any business initiative, data projects can be time-consuming and present surprises. A champion helps maintain project inertia and keeps internal staff focused and motivated. The champion's leadership is essential to keeping projects moving forward and ensuring the organization maintains a proper and accurate course.
Without a Champion, products and deliverables may not meet required business objectives, or internal staff may get distracted and pulled in a different direction. Having a Champion who takes ownership of your data project is essential in ensuring success.
Poor data quality
Poor data quality is the final issue many organizations face when beginning a large data project.
In most organizations, data can be confusing and requires a map to navigate the myriad of levels and pathways to reach what you seek. In some cases, information may be stored in more than one system, even in more than one electronic health record system. Once accessed, data is often complex, static, and not user-friendly. Not only that, but when not properly combined, specific sources seem to have no connection or purpose. These valuable data sets fall by the wayside and fail to be utilized.
Good data is more than just collecting as many data points as possible. Of course, having many data sources is helpful, but data is just numbers without specific goals and targets.
The Pinnacle platform extracts and organizes Electronic Health Record data to specifically address behavioral health agencies' needs. The warehouse allows your organization to accurately capture productivity and utilization across services, including services with multiple providers and clients.
It also can assist in tracking risk factors and outcome measures for population health management. The data warehouse will also allow you to integrate data from payroll, accounting, HR, legacy systems, and other client care systems, creating a central source of agency data that is user-friendly and reliable.
Wrapping it Up
Your organization deserves data that works for you so that when it comes to your large data projects, you don’t fall into the 60% that fail or don’t make it into production. Set yourself up for success by analyzing the common mistakes made by others and working to prevent them.
Pinnacle can assist you in overcoming and preventing each common mistake made in large data projects, as well as help your organization fully utilize its data, increase productivity, improve patient health outcomes, and much more. We will be there to help you during each step.
Let us bring our near half-century of combined experience in data warehousing, reporting, analytics, and business intelligence for behavioral health to your organization.
Set yourself up for success with Pinnacle Health Informatics. Schedule a free chat with us to discover how we can help!