“What do you believe are the most effective ways to improve data quality?” was asked during our 2015 and 2016 State of Data surveys. Interestingly opinions have not changed much; nearly 70% of respondents in 2016 believe a clear definition of what data is needed is the best way to improve, while only 13% believe investment can improve the quality of data.
The running theme seems to be that addressing the data quality isn’t really about reviewing and refining the process, it’s more about the actual data that is being used, as well as the human side of it.
Problems and Solutions
A big issue with the quality of the data is that much of it is decentralised data or has unrestricted access within business, meaning there may be no strategy and informed decision making is more difficult.
Several ways to improve on this can be an increased standardisation, governance and security. As one survey respondent suggested:
“We should hand ownership of data over to the business side, guiding and mentoring them on how to establish their own rules and objectives”.
The one simple issue is that there is too much data in order to create a successful and impactful data strategy.
The solutions to this can include scaling the data down and focusing on agility and quality rather than quantity. It can also be good to align the data needed to the business objectives.
From our survey it is clear there is a disconnect between IM functions and business units, when asked “how has you data improved since 2015?” those who answered “extremely poorly” cited the reason to lack of buy-in and understanding from business units.
One survey respondent suggested that to improve this:
“We need to improve our listening skills and actually change our communication so that we talk about value rather than monetisation”.
It is clear that communication is key to changing the culture and improving the relationship between the different departments.
It is clear that quality really is more important than quantity and gathering data on its own is not enough. In order to have the good quality data there must be a framework and context for having it, and from our survey results we’d suggest much of this is due to the disconnect and lack of understanding between IM functions and business units.
Does your business suffer from bad data? What do you believe is the most effective way to improve data quality?
[New Whitepaper] Adrift on a sea of data: The state of data 2016. Download the full results for FREE here
"You need to have a combination: defect prevention on a process level and defined information stewardship (responsibility for quality information). Without this it is hard to keep information as much as high quality. The current companies are mostly information based organization (by definition of Peter Drucker) and “… every employee is responsible for quality of information produced and distributed to other information consuments..”. It is about a company culture …. Cleansing itself is only a cost adding activity increasing the costs of stored data which could not be 100% accurate from a point of business process. It is possible to find examples where data are from a profiling point of view good quality but at process level – say poor generating additional activities that cost companies a lot of money is a total sum." - Milan Kucera