While organisations have embraced data and analytics, a new KPMG survey shows the road is fraught with concern about trust. But KPMG's Anthony Coops says there is a growing appreciation that connectedness and transparency are fundamental pillars to resolving this.
Issues of trust, ethics and return on investment (ROI) in data and analytics (D&A) are challenging organisations in Australia and around the world, with a new KPMG survey finding that less than half of respondents are confident about the insights they are deriving from D&A in areas such as risk, customer behaviour and business operations.
The findings, detailed in Building Trust in Analytics – Breaking the cycle of mistrust in D&A, revealed that Australian respondents are largely on par with their international peers, sharing key concerns around the quality of their data, and how it is being used and valued.
Whilst data quality is often flagged as a basis for a lack of trust – trust in the data sourcing stage is relatively high (even at only 38 percent) and it quickly falls to just 11 percent when it comes to the use and deployment of analytics. Just 10 percent of respondents placed the most trust in their ability to measure the effectiveness of their analytics.
Anthony Coops, Partner, Data & Analytics, KPMG, says while it is common to hear people criticise analytics and blame poor data for this, there is more to the issue than the quality of the data alone.
"The problem is it has become too easy to blame the quality of data," he says. “As organisations increase their use of D&A they are discovering concerns elsewhere, including with the models themselves, such as their inconsistency with other models and resilience."
He says the financial services sector stands out for having the confidence to ask for external validation of their models.
"There are a whole lot of reasons why you should be really careful around your decision making as more and more is being automated and turned into algorithms. Or at least understand the last time someone undertook an impartial review or sense check of consistency of inputs with other similar models," he says.
A deep distinction between confidence in compliance and privacy, versus confidence in ethical use of data, was exposed.
"Many organisations think they meet compliance obligations (44 percent) – but when asked if they believe they excel in privacy and ethical use of D&A, it drops down to 13 percent," Coops says.
There is a significant gap emerging between compliance and having transparency about what data they hold and how it is being used.
"People think of the privacy aspect, but not the ethical use of data issue. Transparency is the key here and without it, trust quality drops away. More and more, we will see a focus on transparency, particularly around the use of customer related data."
A common complaint with D&A initiatives is that organisations do not understand the ROI. As noted earlier, just 10 percent of respondents placed the most trust in their ability to measure the effectiveness of their analytics.
Coops says disappointment with ROI is often due to a misalignment of expectations, and poorly formulated measures of success.
"Organisations that say they are not getting the required ROI possibly need to relook at how they have defined return. They may find they are too focused on a very narrow investment case and not focusing on the portfolio of use cases or indeed, struggle with the upfront investment in time and technology and the impact of stage and gates," Coops says.
A concerning find was that nearly half of respondents said their C-level executives do not fully support their organisation’s data and analytics strategy. One way to combat this is to view D&A projects as a holistic portfolio, rather than solely assessing each individually, Coops says.
"When you do that, you get a greater understanding of the benefits you are obtaining across the board, and the justification. You are more likely to see the links between the impacts of each initiative and be able to explain the commercial returns in business speak."
While fully leveraging the value in your data is a long way off for many organisations, Coops says there are ways they can make noticeable improvements now. In addition to considering the quality of the models and measurement of ROI, engaging the broader ecosystem for assistance can make a difference.
"There are new and emerging ways to better enable your potential for success. People are embracing the ecosystem of startups to help solve data problems, aligning with third party providers, academia, professional services and other organisations," Coops says.
He says as little as 2 or 3 years ago, this embrace of a collaborative approach was rare.
"There was concern that if you told anyone what you are trying to discover you would lose competitive advantage. Some clients, for example, are now making it public and saying, ‘here are my three use cases, how can you help me, can we work together?’"
Coops says another example of working with the ecosystem are local fintech companies enabling the sharing of data between organisations in a highly efficient and secure manner without actually sharing data at all.
“The winners will be those businesses that make it really easy to install the connecting software, with uncompromising security and a focus on connectedness and simplicity.”
He says that as ‘gut feel’ and subjective decision making moves aside for objective data driven insights, it is essential that organisations have trust in their D&A.
“In this uber-connected world where more data flies around us faster than ever before, we cannot risk losing sight of our community and customers’ best interests.”
For more information on the role of data and analytics in organisations, read Building Trust in Analytics – Breaking the cycle of mistrust in D&A.