Jim Ericson and James Taylor presented the results of Decision Management Solutions’ cloud predictive analytics survey this week in the webinar Predictive Analytics in the Cloud 2013 – Opportunities, Trends and the Impact of Big Data. The research methodology included 350 survey responses, with a Web-based survey used for data collection. The survey centered on the areas of pre-packaged cloud-based solutions, cloud-based predictive modeling, and cloud deployment of predictive analytics. You can see a replay of the webinar at this link.
Key takeaways of the study results released during the webinar include the following:
- Customer Analytics (72%), followed by supply chain, business optimization, marketing optimization (57%), risk and fraud (52%), and marketing (58%) are the areas in which respondents reported the strongest interest.
- When the customer analytics responses were analyzed in greater depth they showed most interest in customer satisfaction (50%) followed by customer profitability (34%), customer retention/churn (32%), customer management (30%), and cross-sell/up-sell (26%).
- Adoption was increasingly widespread and growing, with over 90% of respondents reporting that they expected to deploy one or more type of predictive analytics in the cloud solution.
- Industries with the most impact from predictive analytics include retail (13% more than average), Financial Services (12%) and hardware/software (4%). Lagging industries include health care delivery (-9%), insurance -11%) and (surprisingly) telecommunications (-33%). The following graphic illustrates the relative impact of cloud-based predictive analytics applications by industry.
Adoption of Cloud-based Predictive Analytics by Industry
- The most widespread analytics scenarios include prepackaged solutions (52%), cloud-based analytics modeling (47%) and cloud-based analytic embedding of applications (46%). Comparing the 2011 and 2013 surveys showed significant gains in all three categories, with the greatest being in the area of cloud-based analytic modeling. This category increased from 51% in 2011 to 75% in 2013, making it the most likely analytics application respondents are going to implement this year.
Comparison of Analytics Applications Most Likely To Deploy, 2011 versus 2013
- 63% of respondents report that when predictive analytics are tightly integrated into operations using Decision Management, enterprises have the intelligence they need to transform their businesses.
Impact of Predictive Analytics Integration Across The Enterprise
- Data security and privacy (61%) followed by regulatory compliance (50%) are the two most significant concerns respondent companies have regarding predictive analytics adoption in their companies. Compliance has increased as a concern significantly since 2011, probably as more financial services firms are adopting cloud computing for mainstream business strategies.
Concerns of Enterprises Who Are Using Cloud-based Predictive Analytics Today
- Internal cloud deployments (41%) are the most common approach to implementing central cloud platforms, followed by managed vendor clouds (23% and hybrid clouds (23%). Private and managed clouds continue to grow as preferred platforms for cloud-based analytics, as respondents seek greater security and stability of their applications. The continued adoption of private and managed clouds are a direct result of respondents’ concerns regarding data security, stability, reliability and redundancy.
Approach To Cloud Deployment
- The study concludes that structured data is the most prevalent type of data, followed by third party data and unstructured data.
- While there was no widespread impact on results from Big Data, predictive analytics cloud deployments that have a Big Data component are more likely to contribute to a transformative impact on their organizations’ performance. Similarly those with more experience deploying predictive analytics in the cloud were more likely to use Big Data.
- In those predictive analytics cloud deployments already operating or having an impact, social media data from the cloud, voice or other audio data, and image or video data were all much more broadly used as the following graphic illustrates.
Which Data Types Deliver The Most Positive Impact In A Big Data Context