Digital transformation isn’t quick or easy, and in fact many companies are failing in their efforts.
According to Oliver Schabenberger, COO and CTO of SAS, speaking at the SAS Analytics Experience 2019 event in Milan, only 28% of major corporations succeed in their digital transformation efforts.
He says that there are numerous reasons for failure. “Business transformation is almost always seen as a technology problem, but it’s never just a technology problem. It’s a technology, plus people, plus process problem,” he says.
There’s often also a lack of clear strategy and the role of technology in the transformation. He also blames uninspiring leadership, which is often combined with siloed transformation agendas that lack the C-level vision. And lastly, he says there’s an unwillingness in the organisation to adapt.
“Many projects are failing because they’re working in an ecosystem that’s poorly integrated and doesn’t support enterprise applications. There’s also been a tools explosion, individual departments are trying out the flavour of the day (technologies) with no alignment to each other or the rest of the business and that’s creating a mess that needs to be untangled and creates technical debt that has to be pared down.”
To succeed and bring digital transformation to life, he says, we need to integrate technology, process and people. This, he says, brings analytics to life. Both cloud and AI play an important part in digital transformation, he adds.
The shift from intuition-driven decisioning to data driven-decisioning is a heavy lift, he says. “It’s no accident that cloud plays a major role in this.”
Open Model Manager
But failure isn’t limited to the broad play of digital transformation. SAS used the conference as a platform to launch its Open Model Manager, which is targeted at organisations not using SAS products and primarily using open source languages like Python and R.
SAS says Open Model Manager is designed to help organisations working on cutting edge data science efforts, particularly the ‘last mile’ – to put analytics programmes into operation.
“There are an awful lot of these implementations that are failing to see the true business value of many of these programmes; they typically fail in the deployment process,” says Tom Fisher, senior VP of business development at SAS.
“As analytics adoption has increased, data scientists are turning out more and more complex models. Security and data privacy concerns are top of mind when you talk to IT leadership. At the same time, organisations are managing multiple analytical programming languages from open source as well as commercial providers like SAS, which is leading to a gap between the promise of analytics and AI and the reality that organisations see,” says Fisher.
IDC has stated that only 35% of organisations have analytical models fully deployed in production. Of those that are deployed, says Fisher, 90% took over three months to be deployed, and 40% took seven months.
In addition to the SAS Open Model Manager, SAS has launched a global initiative aimed at operationalising the last mile of analytics.
This initiative comprises SAS ModelOps, a packaged offering including SAS technology and advisory services, a ModelOps health check assessment to help understand how to optimise deployments to get the expected performance, a ModelOps framework, which Fisher describes as a ‘DevOps type of handbook’ to help move models through the analytics lifecycle, and education services.
“Deployment is crucial in turning big data into big value,” he adds.
Share