So here are some more sneak peek early results:
- We identified three scenarios – Decisions as a Service, Cloud-based Modeling and Cloud-based Model Deployment – and all of them have become much more widely adopted.
- Pre-packaged Decisions as a Service applications is the most widespread today.
- While lots of respondents are planning to deploy each of the three scenarios, cloud-based modeling is the least likely to be adopted in the future.
- Nearly 2/3 have at least one scenario deployed widely so far and nearly 90% except at least one to be widely deployed in the next few years.
- Despite all the talk of skills shortage the potential for cloud-based solutions to reduce the need for rare skillsets was the only driver that has dropped since 2011.
- Data security and privacy is still #1 as an obstacle for predictive analytics in the cloud.
- That said, there’s less worry about complexity, and more worry about regulatory concerns, than there was in 2011.
- Models are still dominated by data that is static or batch though there was a big increase in the importance of near real-time operational data since 2011.
- When it comes to Big Data:
- Streaming and real-time data still lag in importance.
- While structured data is still critical, respondents see big potential in unstructured data, sensor data, social media data, image/video data and especially voice data in the next few years.
- Curiously unstructured data hasn’t increased in importance since 2011 – all the potential, apparently, is still in the future.
- Private clouds, and managed vendor clouds, are dominant in planning with public and community clouds much less likely to be at the core of future plans.
As for industries, variations between large and small companies and other specifics you will have to tune in to the webinar and read the report to find out. Register for the results webinar and full write up over on Information Management.
