Data Science as a Service (DSaaS) is a comprehensive package of resources and capabilities for data science – thinking of people, algorithms, data and a cloud-based platform to get started – enabling companies that want to become more data-driven to partner with experts in a flexible, organic way that speeds up speed to values.
Speed versus value is crucial here, as too many companies struggle with data science because they are paralyzed by the perception of “bad data”. Every company has its share of bad data, but the myth is also expensive in terms of missed opportunities, that data must be perfect before dealing with data science. DSaaS is great at putting business data to work as it is, and in the process of doing so it cleans it up and makes it ‘AI ready’ so it can quickly deliver value relative to desired business goals.
Another great advantage of DSaaS is that it uses cloud scalability, similar to other software offerings as a service, without turning it into a large, time-consuming IT project. Exploiting the cloud allows companies to do more than ‘stick your finger in the water’ AI and predictive data science, it allows them to jump in right away.
After hearing that the DSaaS solution would provide all the power needed to process billions of bytes of data from his company’s cloud system with minimal IT involvement, the COO of a large investor-owned utility simply said, “Report me.” It did so because DSaaS would allow it to focus on “where” to focus DSaaS cloud processing power, i.e. on its top value creation priorities, rather than on “how” to do so.
Similarly, another data-intensive company saw DSaaS as a great solution to the big data problem they had. After several attempts at an internal attempt to determine satellite weather data, the company found the initiative too expensive and turned to DSaaS instead. What took a year and didn’t deliver quickly turned into a smooth cloud-based solution. Another win for DSaaS and data-based decision making.
DSaaS also gives companies great flexibility in their approach to predictive data science. They do not have to know in advance which projects they will deal with in which order. Sometimes a project discovers something unexpected and needs more resources to maximize benefits; the second time there is no expected value, so let’s move away from that quickly; but other times priorities change and resources need to be directed elsewhere. Everything is fine with DSaaS.
Think of it as a buffet for everything you can eat. DSaaS users can take what they want, in the order they want – DSaaS provides them with a seat at the table.
DSaaS: Transition from ‘Making versus buying’
Traditionally, companies have thought about their AI and data science choices in terms of “making against buying”, but that should no longer be the focus. Predictive data science is not really just the development of a “single” application, but the introduction of a data-based approach in a wide range of operations, making each part better, more efficient and effective. Yes, there may be only one project at the top of the list, but if companies stop there, they look through the wrong end of the telescope.
And while the DSaaS approach adapts to the needs of the organization, it does too supplements internal data science teams. A good DSaaS service provider will bring best practices to internal teams, tested in similar scenarios, making them more effective. Having a DSaaS partner can also help a company “divide and rule” as needed. For example, if a DSaaS provider already has a strong workaround for a problematic data problem, the provider’s team can take the lead. In another area, where a completely new solution is needed, a DSaaS partner may be developing it together with an internal team.
That flexibility goes a long way. Businesses appreciate that DSaaS provides them with a powerful and proven way to become more data-driven, while enabling them to continue at a pace and scope that suits their business needs – and turn around quickly when those needs change.
about the author
Tom Martin leads the data science team at E Source and is a key partner to the company’s clients in demonstrating how data science solutions can help address their challenges. Prior to that role, he was CEO of Data Science at TROVE, and led the implementation of new technology and analytics at PG&E to reduce the company’s operating costs, improve security, and increase network reliability. Tom was recently named Top 25 Thought Leader of the Year by the Utility Analytics Institute.
Sign up for the free insideBIGDATA newsletter.
Join us on Twitter: @ InsideBigData1 – https://twitter.com/InsideBigData1