Top 2017 Business Analytics Trends


Just a decade ago business analytics wasn’t really a reality for most businesses. Tabular and printed reports constituted the analytics world for most business users. Today we have several threads permeating business analytics, including traditional descriptive and diagnostic analysis using business intelligence platforms, visual analytics which support discovery and exploration, machine learning enabling predictive analytics, prescriptive analytics which tell ‘how’ as opposed to ‘what’ should be done, and of course artificial intelligence, which not only brings new possibilities for decision automation, but will affect all the other types of analytical methods. So with this background, here is a forecast for next year, in order of importance:

  1. Analytics become part of the business process. Today we see two separate worlds in many organizations – production systems and business analytics. These will merge, meaning analytics will become part of the production environment through the embedding of analysis into production applications. Superficially this might mean embedding charts and graphs into applications, but we will also see the embedding of predictive models and other forms of analysis. These might be developed internally, or be cloud services which are accessed using an API.
  2. AI and machine learning enabled business intelligence. Descriptive, diagnostic, prescriptive and predictive analytics involve much repetitive work. This is a prime candidate for automation through the use of AI. Data preparation, feature selection, chart selection and formatting, analysis quality checking and other activities will be speeded up by the application of AI techniques. A forerunner of this is BeyondCore, which has now been snapped up by Salesforce.
  3. Less emphasis on big data. Some businesses need to store very large amounts of data. However when it comes to analysis, it is often the case that smaller is safer – for several reasons. Endlessly combining the ever increasing numbers of data features we store, means that analysts will nearly always find what they are looking for, and that it will often be nothing but random noise making itself look respectable. Several books are now available that explore the dangers of big data analytics, the most readable being Signal by Stephen Few.
  4. More concern with analysis quality. It really isn’t enough to construct endless numbers of charts and graphs and simply pick out the ones that look interesting. We need more objective means of assessing the quality of our analysis. Unless this is performed we simply follow our biases and end up with erroneous analysis. AI and machine learning will help with analysis quality.
  5. Direct feeds from analysis to the environment. Analytics is not a world unto itself, and needs to communicate with other devices. Sisense has pioneered the liberation of analytics from the computer display by integrating with Amazon Alexa and other devices. Feeding analysis out to IoT devices and other hardware will make it more widely available, and more easily digested.
  6. Analysis can enable revenue generating products. Large businesses have access to data and analytical power that enables them to create chargeable services for partners, suppliers, and possibly even customers. Groups of businesses will also start to federate their data, creating information services that provide unique insights into their markets – insights that other companies will often pay for. GoodData is a prime example of a supplier operating in this space.
  7. The emergence of AI driven analytical services that transform and automate many business activities. Current examples include Adgorithm, an AI driven media marketing service that continually optimizes placement, timing, ad wording and channel use, to significantly increase lead generation. Another example is Celonis, providing an AI driven platform that performs process mining to predict process performance. In both cases, considerable manual effort has become automated.

The overwhelming theme in this analysis of business analytics trends is the integration of business analytics with the production environment, and integration between different forms of analysis. The other main theme is the effect AI will have on analysis productivity and accuracy. We could hear some big data analytics horror stories leak out, although businesses typically work very hard to keep such things quiet. Even so, we desperately need help to ensure that analysis is rigorous and well founded – which in the main it is not. AI will help with this, and may eventually completely automate it.

2017 will be a turning point. Having experience half a decade of visual analytics irrational exuberance, we will see that visual analysis is just part of the story, and a relatively small part at that.