Managing Data to Deliver Quality, Meaningful Analytics and Business Benefits

Data quality is critical to meaningful data analytics, and data analytics are key to business strategy – but with increased volumes and complexity of data coming into organisations, how can they ensure data is of the required quality and analytics are adding value to the business?

The session was moderated by Garry Manser, senior manager of master and reference data and data quality at BT. Panellists included Mark Wilson, head of data governance at Handelsbanken; Leyre Murillo Villar, CDO and data control lead at BNP Paribas; and Peter Moss, CEO of the SmartStream Reference Data Utility (RDU).

Discussion points included how to use data dictionaries and business glossaries to achieve consistency and help business units understand different types of data, the best way of creating feedback loops between data management and the business, and whether DataOps could help or hinder data management processes.

One concern is that there is simply too much data, and this has the potential to impact quality, which in turn will impact the results of analytics. Without quality standards, data can be less than useful, no matter how much of it you have. To resolve this problem, the first step must be for business and IT functions to work together to ensure clean data that is fit for purpose. This is where feedback loops can be helpful. Does your business understand the data issues? How do you communicate and prioritise problems? Who is the owner of the data? Who will fix it if there are errors? Who will develop indicators and monitor data quality dashboards? All these questions must be answered before an effective data analytics strategy can be put in place.

For example, data is often captured in silos by systems built to serve specific purposes and which were never intended to be used more broadly. With a view of data as a strategic asset, these systems must be joined together to unlock their value, and that poses a further challenge. “Data management has to be embedded within the organisation,” said a speaker. “There is no point if all your capabilities sit with consultants outside the enterprise.”

“Agility is also important,” noted another panellist. “You cannot wait until your data is perfect to start data analytics. You have to be able to demonstrate the business need and the business benefits of investment in good tools and technology.”

The business need includes ensuring data accuracy through real-time monitoring and creating daily processes that are implemented across the organisation. “If data is being used, mistakes are less likely,” explained a panellist. “Quality checks should be based on the tangible use of data. You first have to define what is ‘right’ and then automate as much as possible. This is the only way to deal with the current volume of data.”

Summing up, the panel offered some tips for data practitioners seeking to upgrade data management processes and run effective analytics to deliver business benefits:

1. Assess the maturity of your organisation – where are you in terms of data management, and where do you want to be?
2. Turn things upside down – identify your key business problem first, then work backwards from the analytics outcome you want and the processes you need to get there.
3. Start small – you won’t get 100% of your data correct first time, so identify key priorities and business needs, and implement the changes that will make the biggest impact.

Communicate – update your systems to ensure you speak the same language across the enterprise and ensure you bridge the gap between business and IT functions.