Trust in data and analytics is a non-negotiable priority
Most business leaders say they lack confidence in their ability to measure the effectiveness and impact of data and analytics, says report
Latest News »
- After five decades in politics, President Pranab Mukherjee says ‘Nomoshkar’ to private life
- Opposition boycotts UP Assembly, holds parallel ‘House’
- IT start-up employees are entitled to labour law benefits, says government
- 20 dead, 30 injured in blast near CM’s residence in Pakistan’s Lahore
- Reliance Capital gets shareholders’ nod for demerger of mortgage arm
Most business leaders today believe in the value of using data and analytics (D&A) throughout their organizations, but say they lack confidence in their ability to measure the effectiveness and impact of D&A, and mistrust the analytics used to help drive decision-making, according to a new survey by KPMG International.
For the report, Building Trust in Analytics, KPMG commissioned Forrester Consulting to survey 2,165 respondents from 10 countries to identify the areas in which businesses are using D&A, and to what extent they lack trust in their D&A models and processes to drive decision-making and desired outcomes.
The report shares insights and recommendations on suggested processes, practices and governance for building trust in D&A using KPMG’s four anchors of trust—a framework for assessing quality, effectiveness, integrity and resilience.
Most businesses, the survey shows, use D&A tools to analyse existing customers (50%), find new customers (48%) and develop new products and services (47%).
Yet, executives do not trust that they are managing their D&A processes effectively to generate desired outcomes and lack the necessary measures to assess the efficacy of those models.
Speaking on the India findings, Prashant Yadav, partner, Analytics, KPMG India said, “India is characterized by its diversity and the same is apparent in the variability of analytics maturity and trust in Indian organizations. KPMG research on building trust in analytics shows that Indian organizations on an average rate themselves marginally higher than global averages on anchors of trust (quality, effectiveness, integrity and resilience) but have a huge gap compared to average scores for geographies such as US and Brazil.
“Indian organizations are more concerned with quality of data than the other anchors such as effectiveness, integrity and resilience. As Indian organizations start using more sophisticated analytics which employs large volumes and sources of data and the complexity of the algorithms becomes increasingly difficult to grasp, they will need to raise the bar on all anchors of trust.
“Currently, the Indian regulatory framework and consumer expectations on matters related to trust in analytics are not very evolved. However, these are likely to evolve fast as analytics becomes more pervasive. It will take just a couple of missteps for significant attention to be focused on this topic by regulators and customers.
“Organizations would be able to insulate themselves from these developments by looking at best practices and regulatory frameworks of developed geographies and design to achieve those standards proactively.
“Recommendations in the KPMG report are a universally applicable framework and apply equally to Indian organizations. Adoption of analytics in India is expected to increase exponentially, and setting the right framework and standards in the beginning will help lay a solid foundation for quality, effectiveness, integrity and resilience of data and analytics.”
Low levels of trust
Just under half of the respondents are very confident about the insights they are deriving from D&A in the areas of risk and security (43%), for customer insight (38%), and only one-third are very confident about their insights around business operations (34%).
These low levels of trust may originate at the top and filter down through the organization, the survey data suggests.
Nearly half of the respondents report that their C-level executives do not fully support their organization’s data and analytics strategy. This low level of confidence points to a lack of trust in the insights generated by D&A, which may be due to D&A’s inherent complexity.
A closer look at the analytics lifecycle reveals gaps in trust. Trust is highest at the beginning of the lifecycle—data sourcing—and drops significantly thereafter.
According to the findings, 38% of respondents have the most trust in data sourcing, which is determining which data is relevant for analysis. Twenty-one per cent have the most trust in the second stage, analysis and/or modelling; and 19% have the most trust in the third phase, data preparation and blending.
Trust slides dramatically at the fourth and fifth stages of the lifecycle. Only 11% have the most trust in using/deploying analytics and 10% said the same about measuring the effectiveness of their analytics efforts.
To assess where the trust gaps may be within an organization’s analytics model, respondents rated how well their processes aligned and performed against the capabilities outlined under four anchors of D&A: quality, effectiveness, integrity and resilience.
Quality: Ensuring inputs and development processes for D&A meet quality standards appropriate for the context in which the analytics will be used.
Key finding: While data sourcing was cited as the stage of the analytics lifecycle that survey respondents say they trust most, only 10% said their organizations excelled across all areas in developing and managing D&A.
Effectiveness: Outputs of models work as intended and deliver value.
Key finding: Less than a fifth (16%) of respondents excel in ensuring the accuracy of models they produce.
Integrity: Acceptable use of D&A, including compliance with regulations and laws such as data privacy and ethical issues around D&A use.
Key finding: With the exception of D&A regulatory compliance, where respondents say they perform strongest, they fall well below in achieving excellence in the areas of ethics and privacy with respect to managing trusted analytics. Only 13% perform well in all areas of privacy and ethical use of data and analytics.
Resilience: Optimization of D&A applications, processes and methodologies for the long term. This includes frameworks for governance, authorizations and security.
Key finding: Only 18% say they have appropriate frameworks in place across all areas of D&A governance.