Analytics Outsourcing: ‘Times are a changin’

February 1, 2017 5:25 pm | Updated 4 years ago.

The 1964 Bob Dylan song ‘times are a changin’ may well be the anthem for the analytics service providers with the word ‘perpetually’ added in for effect.

The change in the analytics world has been rapid and transformational. The analytics space has moved from the era of business intelligence to the era of big data and now to the era of the data-enriched offering. These transformations have been effected in the span of last ten to fifteen years.  The skills, technology, infrastructure and the philosophy attendant to the changing paradigm have changed as well. The writing is clear, the analytics outsourcing service providers will have to keep pace or perish.

Most of the large corporations have already shifted to the third era (analytics for data enriched service offering) so as to leverage analytics for a competitive edge. At Schneider Electric, its Advanced Distribution Management System uses millions of data points streaming from the electricity grid to give engineers a visual analytics of the state of the network. At General Electric, the sensors fitted into the turbines, locomotives, jet engines and medical imaging devices manufactured by GE send in real-time data to enable GE to determine the most effective and efficient interval for servicing of the machines, thereby helping GE to advise its clients on maintenance needs and also assemble engineer teams to affect the servicing efficiently.

Like the first two eras of analytics, this one brings new challenges and opportunities, both for the companies that want to compete on analytics and for the analytics off-shoring entities that supply the data and tools with which to do so. Outsourcing vendors will have to respond to new capabilities, positions, and priorities, to respond the shifting paradigms which are listed as under:

Multiple data types will need to be combined: Analytics outsourcing service providers will need to organize multiple data sets of varying complexity and volume to tease out new insights which will help the frontline service delivery personnel to be more efficient. E.g. bringing in additional data (through fitting a new sensor in the trucks) into the optimization algorithms running in a logistical company might help improve the efficiency of its route networks, lower its fuel costs, and decrease the risk of accidents. Analytics outsourcing providers have to factor for this.

Hybrid Data Environment: The first generation data analytics used the tradition query route from data warehouses as the basis for analysis. The second generation analytics outsourcing service providers focused on Hadoop and NoSQL databases. The current requirement is amalgamating the query approach with Hadoop and more. Data management choices have gone up multifold, however, legacy data formats still remain and they have to be moved in through new processes for sophisticated insights. This would entail analytics outsourcing vendors to take a wholesome view of the enterprise-wide datasets available and bring techniques for amalgamation.

Faster technologies and methods of analysis: New technologies are changing the way data is collected, processed and analyzed, the IoTs has brought about the possibility of avoiding moving mountains of data directly from devices into the cloud, rather processed data is moved, which is handled by software enabled hardware attached to the equipment. Processed data has far fewer data points than raw data thus making real-time data analytics faster and cheaper. Such pre-processing of data will become stable for any true blue analytics outsourcing service provider.

Embedded analytics: Analytics is getting hard coded into business decision making process. E.g. Proctor & Gamble is integrating analytics into the nuts and bolts of managerial decision making through pouring in analytics information from 50 different business spheres into employee dashboards, some of which is real time.  Analytics outsourcing service providers will no longer work with data alone, but will also have to suggest how the data can drive decision-making processes in the nooks and corners of a firm.

Cross-disciplinary data teams: With big data comes big analytics, and the time of the analyst is becoming ever more precious and cannot be wasted on routine things like data management activities. Some of the analytics outsourcing service providers are employing data hackers who specialize in extracting and structuring data for the analyst to work on, while the analyst can concentrate full time on data modeling.

Chief analytics officers: No longer will the analytics outsourcing firms be dealing with the IT manager of their client; many large corporations are creating the position of the chief analytics manager give than data is becoming so important for business. The outsourcing service providers will now be dealing with specialists in data analytics and should brace up for being asked to use more sophisticated tools and provide more complex (and complete) analysis.

Prescriptive analytics: The descriptive analytics mines the data to report about the past, the prescriptive uses the past data to predict the future and the current demand for prescriptive analytics uses models to specify optimal behavior.  The prescriptive analytics is predicted to be the driver of the future. E.g. when should General Electric be scheduling the maintenance of the aircraft engine in use with one of its clients is an example of prescriptive analytics. Analytics outsourcing providers will have to concentrate on prescriptive analytics for future growth.  Prescriptive analytics cannot stand solely on the tools of descriptive and prescriptive analytics, new skills have to be learned and sharpened, and the analytics outsourcing service providers will have to invest resources and time to acquire the new tools.

Analytics on an industrial scale: The new tool in the arsenal of the analytics service providers is machine learning. Analytics can potentially scale up to industrial scale through the use of models based on machine learning. This will help organizations use more granular data for more precise prediction. IBM uses 5000 demand generation models to assess which customer account needs focus. These models are all machine generated with a minor tuning by the analyst. IBM has reported that these models have doubled customer response rate compared to non-statistical market segmentation which was manually done and in many cases using perceptions.  The analytics outsourcing firms which go the machine learning route would surely have clients knocking at their door.

In a nutshell ‘times are a changin’ in the analytics space, analytics outsourcing service providers will have to keep in step with the changing times to remain relevant for their clients.