Representative End User Clients
Representative Automation Clients
Representative Software Clients
I've blogged before about the analytics classification that ARC has developed from Tom Davenport's - Describe, Discover, Predict and Prescribe.
I've covered the Describe and Discover styles in more detail already, so let's talk about Predict and Prescribe. If I had to use a single word to describe the difference between Describe–Discover and Predict-Prescribe, that word would be automation. That is, there is a distinct shift towards using sophisticated algorithms to help make sense of the data.
For the Describe and Discover styles, software support is focused on fairly simple - but labor-intensive - data manipulation. Such as data aggregation, for example. However, gaining insight from the data is still very much dependent on the human eyeball and gray matter. With the Predict and Prescribe styles, that changes. Not only does software perform mundane data manipulation, but it also takes the lion's share of responsibility for identifying trends and patterns in the data.
Here's why. We're all smart people, we can all look at the chart below and spot the trend. It's easy - up and to the right.
Only, in this case we'd be wrong as the more complete chart below shows. The reason we're wrong is because the stock market has so much data, and there are so many complex relationships between that data. It's just too difficult for the unaided human brain to synthesize all the data and spot relevant patents. In these situations, predictive analytics - applying mathematical and statistical algorithms to the data - can help to reveal hidden patterns and insights.
This is a mature technology - it's been used for many years in consumer markets. For example, retailers, banks, and cell phone providers use predictive analytics to drive market segmentation, up selling, and cross-selling. Target infamously used it to target mothers-to-be. Usage is now set to grow in industrial settings - driven by the emergence of the industrial Internet of things (IIoT). For many industrial enterprises, IIoT is the manifestation of big data. And as noted elsewhere, it's the view of ARC that asset lifecycle management will be the bridgehead into IIoT applications for many companies.
Predictive analytics can identify and present hidden patterns in data. Prescriptive analytics is a step beyond, suggesting and recommending appropriate courses of action. In my mind, this class of analytics incorporates older concepts - such as expert systems - as well as more recent innovations such as cognitive computing. Prescriptive analytics solutions:
A great example is provided by the application of IBM Watson to cancer care. This solution ingests 12 million pages of medical text (among other sources) to build and maintain its medical expertise. That expertise is then used to suggest diagnoses and treatments to doctors, along with supporting information. Closer to home for many ARC followers, the same technology is also being explored for oil and gas.
I've covered the Describe and Discover styles in more detail already, so let's talk about Predict and Prescribe. If I had to use a single word to describe the difference between Describe–Discover and Predict-Prescribe, that word would be automation. That is, there is a distinct shift towards using sophisticated algorithms to help make sense of the data.
For the Describe and Discover styles, software support is focused on fairly simple - but labor-intensive - data manipulation. Such as data aggregation, for example. However, gaining insight from the data is still very much dependent on the human eyeball and gray matter. With the Predict and Prescribe styles, that changes. Not only does software perform mundane data manipulation, but it also takes the lion's share of responsibility for identifying trends and patterns in the data.
Here's why. We're all smart people, we can all look at the chart below and spot the trend. It's easy - up and to the right.
Only, in this case we'd be wrong as the more complete chart below shows. The reason we're wrong is because the stock market has so much data, and there are so many complex relationships between that data. It's just too difficult for the unaided human brain to synthesize all the data and spot relevant patents. In these situations, predictive analytics - applying mathematical and statistical algorithms to the data - can help to reveal hidden patterns and insights.
This is a mature technology - it's been used for many years in consumer markets. For example, retailers, banks, and cell phone providers use predictive analytics to drive market segmentation, up selling, and cross-selling. Target infamously used it to target mothers-to-be. Usage is now set to grow in industrial settings - driven by the emergence of the industrial Internet of things (IIoT). For many industrial enterprises, IIoT is the manifestation of big data. And as noted elsewhere, it's the view of ARC that asset lifecycle management will be the bridgehead into IIoT applications for many companies.
Predictive analytics can identify and present hidden patterns in data. Prescriptive analytics is a step beyond, suggesting and recommending appropriate courses of action. In my mind, this class of analytics incorporates older concepts - such as expert systems - as well as more recent innovations such as cognitive computing. Prescriptive analytics solutions:
- Often incorporate unstructured data, such as text or images, as part of the source data used for analysis.
- Provide managers with supporting information along with recommendations and suggestions. That is, they provide context so that the user can understand why each recommendation has been proposed.
A great example is provided by the application of IBM Watson to cancer care. This solution ingests 12 million pages of medical text (among other sources) to build and maintain its medical expertise. That expertise is then used to suggest diagnoses and treatments to doctors, along with supporting information. Closer to home for many ARC followers, the same technology is also being explored for oil and gas.