Profile, test, measure, improve, repeat

Learn from your own experience, and Disney's, by leaving nothing to chance and everything to data.


The more attentive you are in your data profiling and validation processes, and meticulous in your documentation, the bigger investment you will be making in future efficiency.

At Disney theme parks, nothing is taken for granted or left to chance. Especially lines. The company works hard to create a seamless "experience" for its 121 million visitors1 each year. Much of their success can be attributed to the fact that Disney subscribes to the concept of continuous improvement, a core tenet of lean integration. They diligently test and challenge all of their assumptions, learning and changing as they go.

Disney relies on years of data gathered by measuring as many variables as possible and then institutionalizing those results. Therefore, they know when the bottlenecks should happen, which is one reason they rarely do.

Watch and learn

As a team and as an organization, you should be able to apply the same continuous improvement principals to your integration projects and change requests. Start with profiling. Identify common factors and end goals among projects. Compare them with factors in past projects, gauge the success of those past projects, and see how much knowledge you can apply. And before moving forward, test your assumptions.

You will be making a big investment in future efficiency if you are attentive in your data profiling and validation processes. And you will add even more value if your documentation process are efficient and effective. You will have objective data to guide your decisions and estimates when looking at change requests. And you will be continuously improving the process.

Reality check

Change requests differ in terms of scope, complexity, and resource-intensity and generally fall into these four categories:

  1. Requests from business users who can get what they need by configuring an existing report, with a little expert help.
  2. Requests from business users who require help from a business intelligence (BI) administrator or super-user to configure metadata in order to add, for example, a certain metric or attribute.
  3. Requests for additions to the data model or extract, transform, load (ETL) logic.
  4. Requests for a new pull of data or a new subject area to be added to the data warehouse.

In order to gauge how long it will take to fulfill their request, what to prioritize, and what factors you will need to measure, you will need to know:

  • What business users are asking for and why
  • What the proposed change would do and mean for the business
  • How that specific request compares with others in the context of the business’ long and short-term goals

With consistent profiling and testing success and metadata knowledge, you will be able to add automation with a larger degree of confidence. And this, in turn, will increase the speed with which you will be able to move the organization at large toward continuous improvement.

To learn more about continuous improvement's role, read the white paper "The Next-Generation of Data Integration: Transforming Data Chaos into Breakthrough Results."

Related content


7 reasons a seasoned developer is an asset to any big data project

Veteran developers will not face obsolescence if they highlight the value of their experience and knowledge.


Turn to CDC for real-time data

If you are looking to extract value from data, remember that big data is still too slow and unstructured for real time.