CEOs lean more towards distrusting data analysis as decision-making instruments, as they lack faith in their ability to measure added value and accuracy of data analysis reports, according to results of a study performed by KPMG titled ‘Building trust in analytics’.
The study shows that the majority of businesses use data and analytics to observe clients’ habits (50%), identify new clients (48%) and develop new products and services (47%). At the same time, CEOs are uncertain about the usefulness and added value of the data and analysis process, as they lack tools that would help them measure the value of applied models.
«Data analysis more often provides decisions that affect us as individuals, businesses and society as a whole. This is why it is important to focus on ensuring the highest possible trust level for data and its analysis. Organizations that continue investing in data analysis but do not focus as much on usefulness and added value from those investments and accuracy are at risk of making decisions based off imprecise models, which only increases distrust towards results,» – says KPMG Baltics data and analytics service manager Andris Aizpurietis.
He emphasized that it could bring about a situation when businesses may find it harder to compete if they fail to use data effectively. 70% of CEOs already believe use of results provided by data analysis puts at risk the reputation of the business at risk.
Nearly half of respondents admit their high-rank CEOs only partially support their organization’s strategy in the field of data and analytics. Such a low trust level points to the lack of believe in reports in data and analytics. This may be related to the complexity of data and analytics.
«Transparency in relation to data analysis process is key to breaking established prejudices in regards to usual decision-making processes being safer. We have to pull the data analysis process from the black box to provide better understanding about it and help organizations believe in the usefulness of the data analysis process and its added value,» – admits Aizpurietis.
Looking at the life cycle of data analysis it becomes clear that trust is most important at the starting point of acquiring data. After that, however, importance quickly declines. 38% of respondents lean towards trusting data acquisition process in which it is decided which data will be used for future analysis.19% trust the second stage of the process – data compilation and preparation. 21% trust in the third process – analysis and modelling. This fourth (application of results) and fifth (measurement of added value) are trusted by 11% and 10% of respondents respectively.
This decline of trust points to the fact that the main challenges are associated with different data analysis result application, study representatives say.
To assess the main causes behind the lack of trust, respondents were asked to assess the data analysis processes of their organizations in accordance with four aspects: quality, efficiency, integrity and flexibility.
Quality – ensuring data input and processing methods comply with quality standards and the general intended application context.
Study representatives note that data acquisition was admitted as the most trusted stage of analytical life-cycle. Only 10% said their organizations can perform high quality data analysis in all fields.
It is important for the management to ask the question – is data quality appropriate and is a systematic and standardized view on data is ensured?
Efficiency – a solution works as intended and offers added value.
16% of respondents said models are precise. Study representatives say CEOs should ask if information received from data analysis is trustworthy to support decision-making process and if models and analysis work as intended
Integrity – appropriate application of data analysis, including compliance with regulations and ethics in relation to use of application.
Only 13% of respondents admit that data analysis successfully includes all matters associated with data privacy and ethics. In order to ensure integrity it is important to make certain that data and analysis is applied according to the intended goal, representatives say.
Flexibility – long-term optimization of data analysis tools, processes and methodology
Only 18% of respondents admit having appropriate internal processes in all data and analytics management fields. In order to secure flexibility, it is important to make certain that all business opportunities are identified in time and measurement of new product or service results is performed in a timely fashion and that all regular processes are optimized.