Register now for - "How to Create, Govern & Manage Data Products: Practices and Products You Need to Know" - Wednesday, December 13, 11:30 a.m. Eastern Time.

Eckerson Group Predictions for 2017

Award-winning author and New York Times columnist, Tom Friedman, says in his latest book, “Thanks for Being Late”, that we are in an age of acceleration fueled by Moore’s Law that is moving faster than our society’s ability to adapt effectively to the changes.  I’ve been around almost as long as Friedman, and I’d have to say he’s dead on. If you’re not getting dizzy from the pace of change, you aren’t tuned in!  

I started covering the business analytics field in the early 1990s. By the mid-2000s, I thought business analytics had crossed the chasm into the mainstream and was well on its way to a sleepy, commodity market dominated by a handful of data and analytics vendors. Boy, was I wrong! 

Rather than slow down, the business analytics market has quickened. It started when open source invaded the data space, spawning dozens, if not hundreds, of Apache projects in various stages of maturity. It then blossomed with visualization and self-service tools that liberate business users from the IT department and BI search tools that make ad hoc queries possible for all business users.

And now, it has absolutely exploded with artificial intelligence, which has given us everything from smart personal assistants, such as Siri and Alexa, to AI-enabled administrative assistants and robots that can walk, talk, and (almost) think like humans. And now AI is intersecting with the internet of things to give us everything from game-changing smart cars, smart factories, and smart cities to prosaic (or frivolous?) smart refrigerators, smart toothbrushes, and smart diapers (yes, believe it.)

Eckerson Group Experts Predict Changes in 2017

And change has come to Eckerson Group as well. With the new year, Eckerson Group has launched three research practices, each led by a veteran industry expert in the field.

Dave Wells, my long-time colleague at TDWI and its top educator and course designer, covers data management; Stephen Smith, one of the early developers of commercial data mining software, author, and software entrepreneur, runs the data science practice; I run the business analytics practice, continuing my work in this space for the past 20+ years; finally, Henry Eckerson has joined his namesake company as a staff writer and junior researcher and is focusing much of his time tracking advanced analytics and artificial intelligence.

I asked each of these research analysts to submit predictions for 2017 in their fields of inquiry. Here is their response.

Data Management Predictions – By Dave Wells: 

  1. The enterprise data marketplace (EDM) becomes a priority. As companies begin to recognize the undesirable side effects of self-service they are looking for ways to reap self-service benefits without suffering the downside. An Amazon-like data marketplace where analysts can seek datasets, see reviews of others, and select the best-fit datasets for their needs helps to encourage dataset reuse, minimize redundancy, and prevent flawed analysis that results from working with less than ideal data. Data cataloging tools, data curation practices, data preparation technologies, and data services will be combined to create a marketplace for data seekers. EDM returns us to the single-source vision that was once touted as the real benefit of EDW.
  2. Master data management (MDM) is recharged. A resurgence of interest in MDM is powered by challenges of legacy MDM implementations and new capabilities of next-generation MDM technologies. Master Data as a Service becomes a reality as MDM vendors build products to deliver master data solutions in the cloud, bringing an answer to the frequent dissatisfaction with the high cost of maintaining MDM. Intelligent MDM tools reduce the level of effort for data stewards, allowing them to keep pace with the workload of managing reference data and easing the pain of integrating data from 3rd parties and big data sources with enterprise master data.
  3. New kinds of data governance organizations and practices emerge. Long-standing, command-and-control data governance practices fail to meet the challenges of big data and of data democratization. We’ve traditionally exercised governance controls with gates, where explicit permissions are required to enforce compliance with governance policies. Early adopters of new governance practices will implement 3 G’s of governance: Guides to foster good data management practices, Guardrails to prevent unintended data mismanagement, and Gates only when rigorous enforcement is truly required. Governance organizations will also begin to change. While traditional governance roles – owners, stewards, and custodians – will continue to be necessary, they will be complemented with 3 C’s of governing organizations to work with democratization: Curators to actively manage collections of datasets, coaches to foster data literacy and data skills, and community to build collaborative data governance practices.

Data Science Predictions – By Stephen Smith 

  1. Self-service and automated predictive analytics tools will cause some embarrassing mistakes. Business users now have the opportunity to use predictive models but they may not recognize the limits of the models themselves. For instance a model may show increasing product sales when special discounts are provided. Without business oversight the model would be happy to recommend discounting up to 100% of the product’s price in order to ‘sell’ the most product. Expect increased sales of Nassim Taleb’s book “The Black Swan”.
  2. The role of data scientist changes from doer to teacherDespite the revolutionary advances in self-service and automated predictive analytics tools. The reason for this is that the power of the new PA tools will open up huge new opportunities for the use predictive models. Watch the data scientists move from a role of doer to the role of teacher and monitor as business users become empowered by these new self-service and automated tools.
  3. Training will continue to be massively under-budgeted for data science. Many successful companies invest significant portions of their total data science budgets to train data scientists to understand the data and the business needs. Likewise, those companies also make substantial investments from their marketing budgets to train their marketers and other business executives on the basic concepts of data science, model construction and the limitations of predictive models. These budgets will continue to be the first ones cut as data lake and predictive analytics projects fall behind schedule or over budget.

Business Analytics Predictions – By Wayne W. Eckerson

  1. Modern analytic platforms dominate BI. Business intelligence (BI) has evolved from purpose-built tools in the 1990s to BI suites in the 2000s to self-service visualization tools in the 2010s. Going forward, organizations will replace tools and suites with modern analytics platforms that support all modes of BI and all types of users on an open, integrated architecture built using microservices. Developers will use open APIs to build unique applications with embed analytics as well as extend the platform with unique features that others can download from a community marketplace.
  2. Companies start monetizing data in earnest. Business leaders in 2017 will get serious about embedding data and analytics to increase value the value of their products and increase market share or upcharge customers for data analytics usage and generate more revenues. Cloud-based application providers will mine product usage data across their customer base to develop highly targeted recommendations and alerts that further enhance the value of their products and services.
  3. Analysis becomes a data service. Rather than building applications, companies will rent data services from providers who will use artificial intelligence and rules to automate mission critical business functions, such as fraud and anomaly detection, business forecasting, attrition analysis, root cause analysis, and time-series analysis. To use these internet “robots”, companies will simply upload a data set and define what they are looking for.

BONUS Prediction – From Henry H. Eckerson 

Use of natural language generation (NLG) tools will increase in 2017. 

The visualizations, reports, and dashboards that business intelligence (BI) products produce have their limits. Business users still have to do the analysis and interpretation, which can be tedious and time consuming. NLG tools supplement BI platforms by automatically performing the analysis and generating an English language translation of what is significant and meaningful in the data, augmenting the job of the analyst. If businesses want to expedite time to insight and increase efficiency during analysis, NLG tools are the logical next step. Expect to see a number of BI platforms integrate NLG products in the coming year.

All in all, expect the pace of innovation to increase in 2017 as artificial intelligence meets the internet of things and data and analysis pervades all business and consumers products, applications, and services.

Henry H. Eckerson

Henry Eckerson covers business intelligence and analytics at Eckerson Group and has a keen interest in artificial intelligence, deep learning, predictive analytics, and cloud data warehousing. When not researching and...

More About Henry H. Eckerson