How a Trust Gap is Holding Back Analytics

The trend is unquestionably growing: Data and analytics (D&A) is increasingly shaping how companies do business. Complex analytics are delivering better, faster decisions; driving rapid investment across all business sectors; and offering the power to unlock untold value in organizations, from product development and streamlining operations to managing risk and compliance.       

There is no question that companies want more D&A in order to take advantage of those potential benefits. Adoption is growing at increasing speed, while the breadth of usage is also increasing. The expectation of analytics use is expanding, says Matteo Colombo, a principal with KPMG’s Advisory Practice, who points out that over 70% of clients expect their companies to use analytics. The impact of D&A, too, is becoming more and more relevant: “For example, we are used to airlines using analytics to keep us safe in the sky, but there are multiple uses of analytics today that impact our everyday lives.” 

With such power, one would think that companies would completely trust the D&A they invest in. Ironically, however, many don’t trust what D&A tells them, even while they consider it essential. According to KPMG’s new report, Building Trust in Analytics: Breaking the cycle of mistrust in D&A, the vast majority of survey respondents say D&A insights are critical to their business decision-making, but just 38 percent have a high level of confidence in them. And only a third seem to trust the analytics they generate from their own business operations.

Why is trust in data and analytics such an issue for many organizations? The KPMG report shines a light on the trust gap that threatens every organization, even the most mature, and measures and benchmarks the current level of trust in the market.

Grappling with a Gap that Leads to a Cycle of Mistrust

There are multiple reasons for the trust gap, Colombo explains. One is the complexity of the analytics ecosystem, which includes third parties. Another is that analytics are evolving towards more advanced models of deep learning, which tend to be more opaque — it can be hard to really determine how the machine makes a decision.

The third reason is the nature of analytics itself: “Analytics are not deterministic by nature, yielding a single result describing a single outcome, which is what executives are most familiar with,” Colombo explains. “Analytics are probabilistic in nature, giving a distribution of possible outcomes with a confidence level, which is something organizations need to adapt to.”

The study found the trust challenge across the entire cycle of analytics, including the process of ingesting and storing data, modeling the data and generating insight, and finally, generating value from those insights. “We found the level of trust decreased throughout the cycle,” he points out. “That shows us that generating value from insight is probably still the biggest challenge organizations are facing.”

Addressing this trust gap is essential, says Colombo, because it becomes a roadblock to building real value in the business. Even in areas in which analytics has clearly generated more value, including around decisions related to customers and to operations, the study found that the level of trust executives put on those decisions is still very low. “That means that there is inherent risk in the organization in using analytics without addressing this trust gap,” he says.

How to Improve the Level of Trust

According to the study, those that are able to overcome the trust gap quickly “will be the ones that will be better-placed to make faster decisions, more accurately and with much greater confidence.” But how can the organization make sure it is in a position to win in the future through trusted D&A?                                                                  

Whether the level of trust is perceived or actual, organizations should take a systematic approach to trust that spans the lifecycle of analytics and is founded on four key anchors of trust: quality, effectiveness, integrity, and resilience.

Quality is, surprisingly, still a big challenge, says Colombo. “It’s not just about trusting your data, but trusting that you even have the right building blocks for the analytics in your organization and the organization understands how to use this capability.” Key gaps in quality may include the appropriateness of their data sources; the quality of their data sources; the rigor behind analytics methodologies and the consistency of D&A processes and practices. Establishing cross-functional teams, simplifying interconnected analytics and adding rigor can help.                             

Effectiveness is all about whether the analytics models are achieving what they were built for. Do they understand that they work as intended? “We found this was also still problematic in many organizations,” he says. The study found key gaps include the way D&A is used across the organization and the accuracy of models in predicting results — monitoring effectiveness and assessing value can aid in increasing the accuracy of data. 

The third component, integrity, refers to ethical concerns — an example is when a leading digital platform for transport services was tracking iPhone devices after a user took the app off their phone. “That’s an example of use of analytics beyond the boundary of trustworthiness and integrity,” Colombo explains. Boosting transparency and reducing risk can help.

Finally, resilience means the operations around analytics can be sustained — optimizing for the long-term in the face of challenges. Bridging the gap between business and D&A professionals, while monitoring goals, performance and impact can all assist in ensuring D&A is being utilizing properly and there is trust behind the key insights.

To be successful in closing the trust gap, organizations need to address all four anchors continuously, and not just on an ad hoc basis — including assessments, goal alignment, increased internal engagement, expertise building, a focus on transparency and encouraging innovation.  “There needs to be an ongoing process of building trust across those four elements,” he says. The CIO can be of particular help in this process, including the potential of putting together a Center of Excellence (CoE). 

“This is absolutely critical,” he explains. “We found that organizations that start their journey by building a CoE, have a much higher degree of success and they’re actually more effective also in addressing this trust gap.”