Customer Analytics

IBM and SAS Analytics

It is hard to find anyone these days who needs to be sold on analytics. Everyone agrees with the promise of analytics. Yet companies experience frustrations when it comes to realizing its potential. A recent McKinsey study found that “86% of executives say their organizations have been at best only somewhat effective at meeting the primary objective of their data and analytics programs”. SAS’ long experience of building out successful analytics programs gives clarity to what it takes to be successful: in order to maximize the value of analytics, you need to quickly access any relevant data, analyze the data with a broad set of discovery tools to produce insights, and enable the organization to take action – thus sustaining the value by efficiently deploying the analytical insights across your organisation. Such analytics strategy can be expressed as the relationship between its three key elements: Data, Discovery and Deployment.

Benefits can be realised at the intersection of any two of the three elements, e.g. Data & Discovery; however, by considering all three elements together the maximum value can be derived.

Data and Discovery

How does your organisation use SAS for data discovery today? Can your analysts and those in the emerging role of citizen data scientist access all the data and tools they need? Are you happy that the appropriate data governance is in place to control and monitor that access? Are you implementing a “Data Lake” and moving some or all of your data to Hadoop?

By allowing the governed loading of data into a Discovery environment and the provision of controlled access to that data, the duplication and proliferation of data can be controlled and the demand on your IT infrastructure reduced along with the cost of storing the data. Such an approach allows an assessment of how valuable the data may be before deciding to incorporate it in a “production” environment or process.

A key element of this governed process is the provision of discovery, visualisation and even Base SAS coding capability via web and mobile interfaces. It removes the need to manage desktop versions and licenses, and prevents the proliferation of “local” copies of data.

It may very well be that your citizen data scientist, making use of the discovery tools available both within SAS and potentially other technologies including Open Source, have discovered a correlation or anomaly that requires further investigation and classification by an experienced data scientist or statistician. Such an investigation using more sophisticated data mining or statistical tools e.g. SAS Enterprise Miner, may result in the inclusion of new data to refine an existing data model or the creation of an entirely new model to better serve customers with next best offers, upgrades or similar.

Discovery and Deployment

Analysts can build predictive models using a variety of SAS tools that include a rich set of algorithms to analyse structured and unstructured data, drawing from the latest text analytics, data mining or machine learning techniques. What about those users that have developed models using R, how can the performance of those models be compared with those written and running within a SAS environment? Later releases of SAS have the ability to incorporate models written in non-SAS languages, including R, into Model Management processes that evaluate challenger models against existing champions and facilitate the promotion of a challenger if it is proven to perform better than the incumbent champion. With SAS, you can then choose from multiple deployment options to get your champion models embedded into your production systems, where they can produce analytic insights as quickly as possible. And the integrated and automated deployment provided by SAS Scoring Accelerators can boost your model deployment performance.

What about the issue of moving the data required for a model from the data warehouse into the SAS environment as part of a potentially long running batch process that as a final step runs the model and generates a file of scores to be uploaded back to the operational environment? You would like to run the process more than weekly or monthly but the run time is perhaps too long, especially the data integration change data capture to determine what has changed since the last time the process was run.

What if you didn’t need to extract that data or perform that data integration? It is now possible to deploy analytical models directly into the database so the scoring takes place “in database” and the results are instantly available to the operational system.

With this capability it is no longer necessary to extract all the data, store it, run the scoring models and then pass the results to the operational system. The model can be run for a specific scenario or individual case at a point in time. Indeed, this capability can be extended and used in near and real time using Event Stream Processing capability opening up further possibilities to better serve your customers.

Deployment and Data

Of course with these new found capabilities it is important to ensure they are performing as expected.

How is that new model performing? Can the results be used to further retrain the model? Has the expected uplift in sales or customer retention been realized?

Once a model is in a production environment and is being executed to provide answers, the champion model is centrally monitored through a variety of reports based on standardized key performance indicators. When the model performance starts to degrade below the acceptance level – based on the centrally managed model training assets – the model can be quickly re-calibrated or replaced with a new model.

A successful analytics strategy means more than creating a powerfully predictive model; it is about managing each of these life cycle stages holistically for a particular insight or model and across the entire portfolio of models.

IBM Global Business Services (GBS) as a SAS Gold Partner and with its many years of successful analytics project implementations is well placed to assist you with your Analytics Modernization journey; the author would be very pleased to discuss with you the content of the above or any other aspects of your SAS usage.

SAS ( is the leader in analytics. Through innovative analytics, business intelligence and data management software and services, SAS helps customers at more than 80,000 sites make better decisions faster. Since 1976, SAS has been giving customers around the world THE POWER TO KNOW®.

More Customer Analytics stories

C-suite executives are ready to double-down on cloud computing

At a time when the hype around emerging technologies such as autonomous cars and augmented reality seems to be sky-high, C-suite executives are keeping their feet firmly on the ground when it comes to investing in a core area: cloud computing. While it might not seem surprising that a majority of CxOs would prioritize such […]

Continue reading

Four steps can help your organization close the gender gap in leadership

Despite abundant evidence that gender equality in leadership is good for business, an overwhelming majority of organizations say advancing women into leadership roles is not a formal business priority, according to a new study from the IBM Institute for Business Value (IBV). In fact, women hold only 18 percent of senior leadership positions among 2,300 organizations surveyed […]

Continue reading

In the Spotlight: Damian Brown

What is your current role for IBM Talent Acquisition Optimization? I am leading go-to-market sales efforts across Middle East and Africa for TAO, a role which has broadened across our Talent and Engagement (T&E) offering over the last six months.

Continue reading