Dresner Report on Advanced Analytics

September 4, 2016

 

The latest Dresner report on Advanced Analytics is now available and can be requested here. I was surprised reading the report with some unexpected - though highly interesting - conclusions. Some of them are listed below, though I encourage all to request the full report.

 

Executive Summary

  1. Penetration and current use of advanced/predictive analytic tools remains low at just 24 percent and declined slightly year over year. Coupled with future plans, we conclude that APA will not be a widespread practice in the near future. Going forward, large and small organizations are most likely adopters.

  2. Organizations view advanced/predictive analytics as building on existing BI efforts and are generally confident on BI execution and ability to act on data.

  3. Ninety percent of all respondents attach, at minimum, some importance to advanced and predictive analytics, in agreement with earlier studies. Industry (vendor) sentiment toward APA increased in 2016.

  4. Statisticians / data scientists, business intelligence experts, and business analysts are the greatest adopters of advanced and predictive analytics. Hybrid roles are also evident. APA use increased year over year.

  5. Regression models, clustering, textbook statistical functions, and geospatial analysis are the most important analytic user features/functions. Feature interest is increasing. Industry support is mostly in line with expectations.

  6. A range of data preparation features, led by de-duplication, set operations, and complex filtering, are important to users; but interest cooled slightly year over year. Small and large organizations are most interested. Industry support is aligned with user data prep requirements.

  7. Usability features addressing sophisticated advanced/predictive analytic users are almost uniformly important today and over time, led by easy iteration, advanced analytic support, and model iteration. Current industry support for usability features is good and improving.

  8. In-memory analytics and in-database analytics are the most important scalability requirements to respondents, followed distantly by Hadoop and MPP architecture. Sentiment over time is sustained. Industry support for scalability is well in line with user requirements.

  9. Industry suppliers rely upon a combination of native and third-party capabilities and, in some cases, multiple products.

Please reload

tweet this article

Share on Facebook

Highlighted articles

the Rule of 7even

July 14, 2018

1/10
Please reload

Recent articles

July 14, 2018