Research Alert Research Alerts
What is Happening?
Data science – the categorization, extraction and use of knowledge from business data for a wide variety of real-world applications – has tremendous business value. And while that value is core to the digital transformation and future of most enterprises, analysis published in a research report from ISG Insights indicates that it is also rapidly evolving, and therefore difficult to understand, and its use is unevenly distributed depending on enterprise size.
As explained in The Business Value of Data Science, published for clients of ISG Insight’s’ Emerging Technologies and Markets (ETM) research View, user enterprises are already applying data science in automating business processes and decision-making (fraud detection and credit decision systems), delivering faster market- and context- relevant turn-around times (recommender systems), and will ultimately transform current business processes into further digitally operated business workflows.
To date, the business value of data science experienced by early adopters includes the following:
- Better financial results (e.g., increased revenue, lower costs, increased profit)
- Larger market share through faster cycle times and faster decision-making
- Larger share of wallet from more responsive, market/customer customization
These business advantages are encouraging further development and use of data science among the early adopters in more parts of the business, more functions and business workflow uses, and in larger portions of the value chain.
However, its value is much greater than simply altering workflows, operating models and financials. Its uses will alter competitive alignment and value chains as we know these today.
Why is it Happening?
Data science and its uses are still perceived as science fiction by many business and IT leaders, far removed from traditional forms-based information processing, and weekly/daily business intelligence updates and dashboards containing key performance indicators.
However, our ongoing research reveals that business and IT leaders should not be lulled into inaction about data science because it appears to be far off. Data science applications are delivering business value today, and will transform how IT is used to run and operate the business of the enterprise.
Early adopters, mostly among the Global 500, are developing and deploying custom data science applications driven by proprietary datasets. These enterprises are experiencing wildly different results, with some of the early big-data experiments accorded duds, and later uses of business driven tests to be winners. Early uses of data science abound in health care, the life sciences, finance, manufacturing, retail, travel, hospitality, energy, services, public sector, utility, and agriculture/feedstock industries among others in attempts to deliver competitive advantage.
In addition to the early stage big-data projects, new Citizen Data Science platforms are alleviating some of the difficulty in building and delivering data science models and applications, making it possible for more firms to deliver custom data science applications.
Furthermore, Cloud-based SaaS packaged applications with embedded data science capabilities are available for mainstream customer relationship management, marketing and sales, factory floor automation, supply chain management, financial planning, HR and talent management, logistics, distribution, and customer service. Business uses of SaaS-based data science applications will – in time – approach and exceed the capabilities of today’s custom applications. Along with in-house domain knowledge and packaged domain data brokerage services, such pre-packaged approaches may level the playing field for common function and process oriented business uses of data science
Additional uses are moving data science to network edges to deliver more responsive application uses for the Internet of Things (IoT) among others.
Still other commodity applications of data science are resulting in Cloud-delivered data brokerages and a variety of specialized SaaS applications that are industry specific to health care, life sciences, finance, manufacturing, retail, travel, hospitality, energy, services, public sector, utility, and agriculture/feedstock industries among others.
Net Impact
The uses of data science will redefine legacy forms-based business workflows and the uses of business intelligence. It will imbue decision-making with better forward-looking recommendations, while it automates rote and programmatic workflows and business functions. Data science applications will enable nimbler market responsiveness and gain greater customer allegiance, stickiness, share of market and share of wallet. In short: the business value of data science will creep up on most, may not be noticeable unless it is too late, and will be stunning in its overhaul of information processing and the business of IT.
Packaged business applications will provide market-leveling capabilities as IT providers deliver general purpose SaaS applications with embedded data science capabilities, and as more data brokerages form and provide grist for data science applications. We also expect competitive differentiation – longer term – will remain steadfast for the combination of industry specific applications of data science, powered by highly proprietary datasets. This combination will retain market differentiations farther into the future.
This will both drive and result from the expansion of data science use through 2020, although the benefits will not be equally enjoyed. Our research indicates significant differences in data science adoption and utility based on relative enterprise size, with relative percentages of data science adoption through 2020 as follows:
- Global 1000 enterprises: more than 80 percent
- Mid-size enterprises: 40 to 50 percent
- Small businesses: about 20 percent
The imbalance in the early uses of data science applications through 2020 – a larger share of global firms, and smaller share of small businesses – will result in market-responsive advantages for the largest enterprises. The largest will use data science to react faster to customer need and market changes, take a disproportionately larger share of wallet as a result, and improve financial outcomes further. This could reverse some of natural advantages enjoyed by smaller and more nimble firms that can respond to individual customer needs and market shifts more rapidly than is typical for large enterprises; if we are correct in this, we can expect to see a resulting data science applications/data arms race within the next few years.