Managing Data Stewards
The extent to which rules can be used to resolve the conflicts and ambiguities arising from inconsistent and incomplete data is limited. At some point it becomes necessary to turn to data stewards to apply human reasoning and additional insight to the data. There are many factors that determine the amount of work required by data stewards, but in some cases the workload can be significant. In this article, we discuss some of the important considerations when planning and managing a large data stewarding effort.
By Martin Dunn
Breaking down a large number of data stewarding events into smaller groups makes it easier to create and maintain enthusiasm within the data steward team. Just as it’s easier to convince someone to run a quarter mile than a full marathon, data stewards will more readily tackle a series of small groups of records than one long list.
Even in situations where there is a backlog of tens of thousands of data stewarding events, breaking these into smaller groups of a few hundred records allows each steward the satisfaction of starting and finishing a set of records several times a day.
Communicate Policy Clearly
Data stewards are called upon to review situations that cannot be resolved by rules which typically imply that these situations are complex or ambiguous. A downside to human intervention is that people are less consistent than computers and this inconsistency is exaggerated in these complex/ ambiguous situations if the policy for managing these situations is at all unclear.
It is really important that the way data stewards handle their workflows is clearly documented and communicated to all stewards.
There will be situations where the data steward does not have enough information to make a decision. Providing some mechanism to move these records out of their queue allows the steward to continue to make forward progress and does not skew the time-per-record statistics that are so important to managing the stewarding process. Another advantage of a “suspense queue” is that these records can be more thoroughly analyzed to determine why they are more difficult to resolve.
Screen Design for Efficiency
User-friendly screens do not imply pleasing-to-the-eye designs but rather efficient-to-the-hands layouts. Your design must optimize the number of key strokes required for any action and certainly minimize the amount of mouse movements which are time consuming.
Data stewards need to assess the data on the screen and action their decision with minimal fuss. Stewards will not thank you for elaborate fold-out lists and pop up information boxes no matter how pleasing the design.
Data stewards are the best people to comment on the efficiency of MDM rules and they should be encouraged to provide feedback to the data management team. When a steward has to resolve the same situation in the data time and time again they will quickly identify the patterns and can provide use cases to the data management team for further investigation. Sometimes simply adding additional cleansing rules may help resolve hundreds of administrative hours per year.
Data stewards are often the last line of defense for enterprise data quality. There is no safety net for many decisions entrusted to data stewards and therefore it is important to have a solid quality program in place. A mistake made by a data steward will typically only surface as an error in some downstream process. It is worthwhile tracing these data errors back through the MDM process to the data steward. Data stewarding decisions should be traceable to individual stewards as poorly trained or careless stewards will quickly show a deviation from the mean error rate.
In our experience data stewards produce an error rate of much less than 1% of the decisions they make. Poorly trained or careless stewards will typically show a far higher error rate and therefore they can be easily identified and corrected.
Data stewarding is a mundane administrative task that adds time and cost to any MDM initiative. However, it is unavoidable in many circumstances and building a trustworthy, efficient data stewarding process is a key success factor for sustainable data quality. Data stewards provide a key role in the MDM lifecycle and their role should be given the time and attention that it deserves.
Martin Dunn was the co-founder of Delos Technology which developed the MDM technology marketed under the Siperian brand. The Delos MDM technology introduced many MDM concepts that are now widespread within the MDM discipline including a data steward console to adjudicate match results, opt-in synchronization, cell level delta detection and the concept of measuring trust.
Martin is now a partner with Gaine Solutions and continues to advance the techniques by which enterprise Master Data is managed.
Key Questions to Ask During Master Data ConsolidationsTypical master data consolidation starts with combining the operational master records from all the data silos where they exist. The key aspect being, creation of master data indexes to support single view; knowing...
Opt-in SynchronizationNot all operational systems will choose to, or be able to, consume the changes made to master data in an MDM hub. The reasons for being out-of-synchronization may be technical, regulatory, political or economic but at some point it will be...
Changing a Match RuleWhen we are talking to companies about our MDM platform we cover a broad range of topics, from measuring ROI, to more technical questions about the way the software operates. A common technical question is "How do we change a match rule?" Our...
Ready to master data mastering?
Subscribe to our mailing list and we’ll send you courses, insights, product updates, and more. Get to know the ins-and-outs of your Gaine MDX platform, features, and solutions.