Having a complete picture of your medical school’s performance at any time is a critical, but difficult goal to accomplish. Data warehousing is key to solving this problem, but it’s no small feat to build. There are many steps required – from data collection, filtering, sanitizing, and shaping to permissioning, analyzing and visualizing that data. With such a large undertaking, what factors should you consider when deciding whether to build or buy a MedEd data warehouse solution?
It’s no surprise that the determining factor between building versus buying often comes down to feature requirements. When considering building versus buying, first ensure all your stakeholders agree on a prioritized list of requirements. Next, do a thorough comparison of existing solutions.
In general, consider buying when tried and tested platforms exist that provide at least 80% of your needs. Only consider building if you have a large number of requirements that no existing solutions meet.
Having a complete picture of your medical school’s performance at any time is a critical, but difficult goal to accomplish. Data warehousing is key to solving this problem, but it’s no small feat to build. There are many steps required – from data collection, filtering, sanitizing, and shaping to permissioning, analyzing and visualizing that data. With such a large undertaking, what factors should you consider when deciding whether to build or buy a solution?
Development, quality assurance, and testing are labor-intensive and complex processes that require a dedicated team of highly skilled professionals. Take inventory – does your school have the people and data skills to build a warehouse and visualizations? Can your school commit to a dedicated team of at least 1 manager, 1 data scientist, and 3 software engineers for the lifetime of the warehouse? What contingencies are in place to keep this project going if key individuals leave with critical project knowledge?
If you’re unable to secure the development resources necessary to begin developing a data warehouse, consider buying a pre-existing solution or find these services through the expertise of a third party.
Support and maintenance
Your data warehouse is one of the key drivers in understanding your institutional data. If it doesn’t function as expected, stakeholders will resort to manual processes exporting data to perform ad-hoc analysis.
Once your data warehouse is up and running, you’ll require a dedicated team of in-house specialists to keep your critical business operations running smoothly, fix critical issues, and answer pressing questions. Don’t forget to budget for this often-overlooked factor when considering a custom build. Expect to budget 3 or more full-time employees based on the size of your school.
If you don’t have the budget for additional full-time employees, opt for a data warehouse maintained by a third-party solution.
What is your school’s timeline to deliver a solution? Data warehouses come with many complex pieces, requiring more time and effort to develop than might be apparent. Couple this with the sobering fact that 1 in 6 IT projects have schedule overruns of 70% (Harvard Business Review), and it’s easy to see how quickly a project like this can balloon in size. Under ideal circumstances, expect custom-built solutions to take 2 to 5 years to be ready for use. Ensure your budget has significant contingencies for overruns and rework.
If you don’t want to wait years to gain insights from your data warehouse, consider partnering with an existing third-party solution.
Human resources and technology licenses are the biggest costs involved in any IT project.
Expect software developer salaries of $100,000 (U.S. News) and annual licensing costs of $10,000 to $100,000. Factor into your budget that, on average, IT projects come in 30% over-budget, with 1 in 6 being 3 times over budget (Harvard Business Review). It’s easy to see how even a small data warehouse project can cost over $1 million before launch.
If your school has limited financial resources available, invest in a platform from a trusted partner.