“Our data is a valuable asset. There is a robust and structured quality assurance process in place to assure and protect our data assets.”
In my blog explaining the importance of Data Attitude I covered how to implement a number of actions designed to improve the quality of the data captured through improving the action and attitude of those who capture the data.
However, like any process that creates an end product, in this case, data and information are the end products, there needs to be some structure to assuring the quality of this data. We need to have a both quality control at the front end of the collection process and quality assurance at relevant points after the data has been collected.
The data and subsequent information will be used by the company to support business decisions that will affect how time and resources such as the focus of management attention and money will be directed. If the data is not correct, then this could mean the company is not making the best use of these resources and could be losing money through poorly informed investment decisions. The quality of the underlying data matter – a lot!
Work Order Data – a £3M Asset
Collecting data takes time and effort. The activity costs the company money. The table below outlines an example of the typical time and estimated cost of collecting work order data on an offshore installation. You can easily adapt this approach and use it as a template for calculating the value of the data at your site.
| Work Order Process | Typical Data Collected | PM Work Orders | Breakdown and Corrective Work Orders | Cost @ £80/hr labour rate (offshore UK average) |
|---|---|---|---|---|
| Creation | • Equipment Tag • Failure Mode • Type of Failure • Defect description • Risk Ranking | 0 | 15 mins | £27 |
| Review and prioritisation | • Quality check | 0 | 5 mins | £7 |
| Planning | • Est. Labour • Spare Parts • Vendor Services • Plant Condition • Isolation requirements | 0 | 30 mins | £40 |
| Execution | • Actual labour • Actual spare parts | 10 mins | 10 mins | £26 |
| History reporting | • Components replaced • Failure Causes • Remedy • History narrative | 15 mins | 15 mins | £40 |
| History review | • Quality check | 5 mins | 5 mins | £13 |
| Summary | ||||
| Total time per work order | 30 mins | 80 mins | ||
| Total number of work orders per site per year | 10,000 | 2,000 | ||
| Total cost | £400k | £213k | £613k Over 5 years >£3M |
Calculating the Value of Work Order Data
The work order data collection process cost the company around £600k a year. Consider that over 5 years, and this amounts to around £3M. This is a massive amount of money. Besides the cost to build the asset there is also the value that can be extracted from the data in decision support. This will be worth far more to the company as why used in optimising maintenance strategies to prevent equipment failures and optimising maintenance cost to ensure that the maintenance strategy is achieving the best value for money for the maintenance carried out.
Spending this amount of money on a single investment would most likely trigger your procurement QA/QC process because the company’s management would want assurance it was getting value for money and receiving a quality end product.
This is why it is important to consider the company work order history database as an asset. The asset comprises all the individual work orders. The asset has a value that needs quality assurance and quality control measures in place to provide the end-users with the confidence that the asset can be used effectively.
This means that attention needs to be paid to each and every individual work order. It must be accurate and complete. The more inaccurate and incomplete work orders there are, the lower the quality of the history asset will be.
Furthermore, work order data is something that most companies collect anyway as part of their work order process flow. It’s not a case of deciding not to spend money collecting the data. If poor quality data is collected without assuring its quality, it is wasted time and cost. The data will be collected anyway but may never be used or usable to support business decisions because it isn’t correct or off sufficient quality to be utilised. The phrase Garbage In – Garbage Out may be a cliché but it’s 100% true.
Quality Assurance vs Quality Control
The management of the company must develop a level of confidence in the quality of the output generated by the data capture processes. To provide this level of confidence the process must have quality assurance and control checkpoints built into it.
| Quality Assurance | Quality Control | |
|---|---|---|
| Definition | A set of activities designed to ensure quality in the process that collects the data and resulting information. | A set of activities designed to ensure the quality of data. The focus of these activities is the identification of defects in the actual data records, i.e., incorrect work order information. |
| Focus/Goal | Defect elimination in the manual data collection process. QA is proactive. | Defect identification and correction in completed data records. QC is reactive. |
| How | Establish a high-quality data collection and management system. This includes adequate audits of the system for conformance to the procedures and processes. | Identification and then systematically eliminating sources of data quality issues through manual and automated quality checks. |
| What | Clearly documented processes Training Fit for purpose computer systems | Clearly documented processes Training Fit for purpose computer systems |
| Who (responsibly) | Everyone who is involved in the collection of data | A specific role identified in the process for the collection of a type of data, i.e., Maintenance Supervisors for work order data or Operation Supervisors for equipment availability data. |
| As a Tool | Managerial tool | Corrective tool |
| Orientation | Process-oriented | End product-oriented, i.e., Work Orders, Availability records or Production Loss records. |
Key Maintenance & Reliability Manual Data Collection Processes
In post on Data Model, I cover the maintenance and reliability data model. This model describes the data required to manage the effectiveness of the M&R process. Below I’ve covered the specific QA and QC elements I recommend are in place for the processes that generated the data in each process.
| Data Generation Process | What | Who |
|---|---|---|
| Availability process | QA – Training on understanding the difference between planned and unplanned downtime. QC – Validation and approval of availability records each day, including the reason for the downtime | - Operators - Operations Supervisors |
| Condition Monitoring Measurements | QA – Training and periodic assessment of correct sampling and measurement techniques and tools. QC – Review of findings | - Operators - Condition Monitoring Technicians - Reliability Engineer |
| Work order process | QA – Training on the CMMS and expectations of the data required for a high-quality work order. QA – Configuration of failure codes in the CMMS to ensure that data has consistent descriptions. QC – Implementation of a work order review meeting to review the quality of the data entered into work request prior to approving as work orders. QA – Training on the requirements of a good quality work order history report, including an overview of what the data will be used for. QC – Review of work order actual labour, actual spare parts used, failure codes and history narrative priority to committing to history. | - Technicians and Operators - CMMS Lead - MTL/PTL - Supervisors |
| Production loss recording process | QA – Training on the requirements of high-quality unplanned loss record. QA – Provide training on how to carry out an effective RCFA/Why 5s. QC – Review an audit of 5-10% of completed RFCAs each month. | - Production Team Leaders - Production Supervisors - Operators |
