The short answer is “yes”. When you consider why you should consider metric reporting as an important part of your oil and gas data management strategy, the answer is simple: Data management is hard. There are a lot of moving parts to make up a comprehensive oil and gas data management strategy within an organization.
With that understood, the issue becomes, what steps can be put in place ahead of rolling out a data management strategy that can make a difficult process a little easier? That’s where metric reporting comes into play.
What is metric reporting?
A good place to start is the dictionary definition of “metric” as “a standard of measurement by which efficiency, performance, progress, or quality of a plan, process or product can be assessed.”
This definition sums up metric reporting in oil and gas data management. We are using a defined set of quantifiable tests to measure:
- How much a new system is used
- How much data is loaded through that system
- Whether the data is entered correctly
- How data quality changes over time
- How efficiently the system loads data
- Where any bottlenecks are located
To determine the above, metrics should be collected at regular intervals, either daily or weekly, and then rolled up either monthly or quarterly so that a reporting/business intelligence tool can produce detailed reports that show management in a quantifiable way where the money they invested in a new system goes.
How to collect or derive these metrics
Metric reporting, in practice, depends upon the system you are measuring. Some systems have built-in quality control tools; others don’t. Some of my oil and gas clients use a third-party QC tool or develop an in-house QC tool. Any of these solutions can be used to measure metrics against your application and, more importantly, against your data.
As applied to a data management strategy, metrics should fall into one or more of the following categories:
- Quality. Does the data meet a low, medium, high or fail quality standard? This category is really about data governance at the enterprise level and whether your organization has a data governance team that will determine the data quality parameters before the data management strategy is implemented. While it is understood that some aspects related to data quality may change, and even improve, during development and initial implementation, the key is to understand that the main definitions must be made before the implementation of the data management strategy.
- Consistency. Is the data quality consistent? Determining consistency in oil and gas data management can be as simple as checking that wells have completions associated with them or that logs have curves associated with them.
- Correctness. How correct is the data? Data correctness is linked to data quality (i.e., what percentage of your data is of low quality?). Correctness can also be determined by checking if a particular attribute’s value falls within certain boundaries. For example, does a log top depth fall between the top and bottom depth of the well or completion?
- Currency. This addresses key questions such as: How up to date is the data? Are you loading data daily? When was the last date you loaded data?
It should be noted that prior to implementing a new data management strategy, a current state of the data should be captured via data quality checks. This can show users and more importantly management how their data currently looks, and can also be used as a way to measure the quality of the data management strategy.
Data quality dimensions
Metrics and data quality are intrinsically linked. To better understand how metrics are important, it is also important to define some dimensions around data quality:
- Completeness
- Are all required attributes present?
- Are relevant optional data present?
- Consistency
- Is there consistent data within a record?
- Is there consistent data between records within the same table?
- Is there consistent data between records and different tables?
- Are there “reasonable” values for attributes?
- Timeliness
- The amount of time it takes to take an external file and load it into the system.
- Uniqueness
- There are no duplicate records.
- How do you define a duplicate record?
- There are no duplicate records.
- Accuracy
- All data is valid.
- Error detection is part the system.
- Data and application are fit for their intended purpose.
Final thoughts on metric reporting
Although these categories are fairly easy things to define on paper, it takes the cooperation of many different players in your data management strategy to make this work. IT and the business have to talk to one another, and they also have to communicate effectively with the software vendor to ascertain what is currently possible in a system and what will need further development.
As you can see, metric reporting is not a simple project to implement. But it is an effective approach to lay the foundation for your oil and gas data management strategy. And once it is implemented, over time you will see improvements in your data, and your data management strategy will be the stronger for it.