Maksim Kabakou - stock.adobe.com
Data teams struggle to maximise business benefits
Many data scientists are having a hard time moving data science initiatives beyond proof-of-concept projects
Businesses are failing to understand the benefits of data science initiatives, a recent study suggests.
Almost one-third of the companies surveyed for the DataIQ and Domino Data Lab How to scale data science research admit they do not know what business benefit they can expect from data science. More than half see data science as only driving nominal gains of less than 5% impact on annual revenue, while 4.9% of organisations view data science as a cost centre.
Just over two-thirds of the 101 organisations surveyed say they have established a centralised data science function to support the business. This leads to a gap in understanding between what the business wants and what the data science team thinks is required, with one in eight (12.8%) failing to create compelling uses for data science to involve real-world business problems.
Worryingly, 39.5% of the businesses surveyed say they face the challenge of weak understanding and support from the business. This seems to get worse for organisations that consider themselves more advanced in their level of data science maturity.
According to DataIQ and Domino Data Lab, organisations that consider themselves in the “advanced” and “reaching maturity” states report higher-than-average levels of conflict (52.4% and 50%, respectively), making this their main challenge.
In the study, DataIQ and Domino Data Lab warn: “What this approach risks is limiting the scope of projects being undertaken to proofs of concept because data science is not sufficiently embedded into the business to gain full sponsorship or to support truly transformational projects. For that, there needs to be close and regular dialogue with lines of business to identify scalable opportunities and increase understanding of the business among practitioners.”
Another factor identified by survey respondents is lack of IT support, which continues to be a significant constraint in some organisations. One-third of data scientists admit they are in conflict with the IT department.
Domino Data Lab and DataIQ suggest that if this relationship could be improved, it might also resolve the problem with model deployment identified as the third biggest challenge by 34.9% of the data scientists surveyed. “When tackling the issues identified with data science staff, some soft skills training might well prove to be of value here,” they note in the report.
With four out of 10 organisations that took part in the study admitting they need to ensure requirements are better captured from stakeholders, the report recommends organisations to strive to improve the data culture to allow data science to flourish.
Domino Data Lab and DataIQ urge IT and business leaders to develop “active listening” skills and project management expertise among data scientists to ensure they are exploring fully what is needed before embarking on projects.
The business side also needs to be educated around core concepts to demonstrate the art of the possible with data science, they add.
Read more about data culture
- In an interview, Jennifer Redmon, chief data and analytics evangelist at Cisco, talks about how she helps organisations enable their employees to work with data.
- Data has become one of the most valuable assets in the enterprise. IT teams must make changes – both culturally and technically – to ensure their strategy reflects that.