So how can the accuracy, integrity and dependencies such as timeliness of data records be maintained?
For all data there is an individual or a group of people who offer a good source of expertise for that piece of information, or more than one person who will have an informed opinion. These are the people who can reliably be expected to have their fingers on one or more particular pulse or have specialist knowledge that can be applied. The key challenges are to first identify who these people are and secondly to link them as efficiently and effectively to the record stores.
The right people
Getting the right individuals involved has less to do with their skills and capabilities, and far more to do with their roles and position in various business processes. Certain individuals are in the right place at the right time to understand and gather information as it happens. These are the ones best placed to enter and edit records, but in all too many business processes there is a dislocation between the gathering of information, and the subsequent keying in to the system. There are many examples that intersect different industries:
- Field engineers filling in forms by hand to be keyed in at the depot when it would be more efficient for the engineer to fill in the information directly while mobile.
- Telephone operators or ‘human modems' who enter user account information or orders that could more rapidly be handled with online services and forms, backed up by for those with no internet access rather than completely dependent upon human re-keying.
- Factory floor workers exchanging paper or simple electronic requests and bill of parts with supplies and stores managers, where an integrated approach would allow the individuals in the right place to directly key in distribution, usage, shrinkage and breakages. This would boost accuracy and timeliness.
Editorial control
It might seem that the collaborative input of several people will lead to a higher quality output, but it still needs some level of management or control to ensure quality. Much has been made of recent advances in the use of Web2.0 technologies for user-generated content, and encouraging unstructured multiple person involvement through tools such as Wikis. While these tools have their place, there are times when they will lead to poor data quality. The free for all nature of Wikis for example does not guarantee that those with critical knowledge or information are engaged at the right point in the process.
There is also a need to support human intervention through sign-off procedures to ensure checks on quality control have been made where compliance is necessary. Some procedures and information records are so complex that hierarchical or sophisticated networks have to be developed and put in place to provide the right structure for gathering the data. Such structure is unlikely to emerge from consumer-oriented ‘wisdom of the crowd' Web 2.0 tools. Successful high quality collaboration depends on the right initial information/data model and applying editorial quality checks on the end results as well as getting the active involvement of as many worthwhile contributors as possible.
Integrated flow
Even with the right people involved, getting the information into the ‘right' place or application is by no means simple either. In times gone by, prior to interoperable cut and paste, or in pre-graphical text-based full screen applications, many users would typically read data off the screen, write it down and then re-key it in. Not only is this inefficient, it leads to the easy introduction of errors and can also open up security risks. Good integration practice requires a focus on information flow across the business processes to support users in the business's quest for data accuracy.
The critical issue about quality is that it is more noticeable by its absence than presence. The problem with data records is that inaccuracies, data entry, or editing errors and poor checking are rarely noticeable sufficiently early. Too often, once digitally stored or presented, data is assumed to be correct. When poor quality or errors are discovered it causes a collapse in confidence that is proportional to the importance of the data to the user.
No one likes to be let down by bad information. While checking, weeding and correcting records are all possible, the best approach is to aim to get the information right the first time, by supporting the right people at the right time and place in the business process.
