By using website you agree to our use of cookies as described in our cookie policy. Learn More


Embedding Advanced Analytics

To get the most value and greatest insight from your analytics, embed your analytics into your business processes.

As advanced analytics begins to hit the mainstream, so, too, does the move to embed analytics into the business process. When you embed analytics, you actually insert it into the operational systems that are part of a process. For example, a statistician might build a predictive model that predicts what customers will purchase. If the model is not embedded into a business process, then it provides insight, and it may be acted upon manually, but it may not provide significant value.

If that model is embedded into systems that feed the call center, however, the output of the model can be used by a call center agent as part of a business process -- say, to upsell or cross sell a customer who is on the line. Based on the behavior of other customers with a specific profile, a message to the agent might appear on the agent's screen when the customer calls in. The agent doesn't need to know how the model works, just how to work the offer.

Another example is analytics embedded into a system at the point of sale. An algorithm (for recommendations, for example) might be running behind the scenes. Based on what the customer is buying, the recommendation system might provide some coupons if the customer is in a physical location. If the customer is online, it might suggest some other items a customer might buy.

TDWI Research indicates that companies are starting to embrace operationalizing and even embedding advanced analytics into their systems and processes. In our 2014 Predictive Analytics Best Practices Survey, for instance, at least a third of those respondents who were using predictive modeling were operationalizing it as part of a business process. Some of this may be part of a manual process today, but more companies will embed the analytics into systems as part of a workflow in the next few years.

There are numerous reasons why this is important:

  • It makes advanced analytics more consumable. When analytics are embedded into business systems, the end result is that analytics become more consumable, which means that more people can make use of sophisticated analytics output. I've written about this as a kind of multiplier effect in previous articles and blog postings. One person might build the model, but the output of the model is available to far more people. That means that insightful and valuable output from predictive models can be utilized across the organization.

  • It makes analytics actionable. Analysis without action is not that useful. When you embed analytics into a process, it means that the output can be actionable. When used together with business rules and operationalizing it, action can be taken. This might be done semi-automatically or automatically That can be a big advantage for companies. For instance, a fraud application can utilize embedded analytics to detect probable fraud based on the characteristics of the transaction and automatically route the instance to a customer for verification or a special investigation unit. A preventive maintenance application can monitor assets for issues based on past patterns or rules and trigger alerts that can improve performance and save money.

  • It makes analytics more valuable. As analytics become part of a workflow, it can provide top- and bottom-line impact. The examples above support this. Automating analytics can reduce costs associated with manually trying to deal with the output. The analytics can be utilized by more parties to drive benefits. Operationalizing and embedding analytics makes organizations more productive. Many organizations doing this feel that it is a competitive differentiator.

  • When analytics are embedded into business systems and processes, the analytics can become more prescriptive. Prescriptive analytics is a form of advanced analytics that utilizes predictive analytics and suggests options based on predictive output. Some people consider this the next evolution of predictive analytics (where predictive was the next stage in sophistication from descriptive analytics). Predictive analytics helps determine what is going to happen; prescriptive analytics can help you take action. A recommender system we described is one example of a prescriptive analytics system. Examples in varying shapes and level of complexity can be found across all industries.

Of course, gaining value from any kind of analytics requires a solid infrastructure and the organization to support it. Learn more about these topics at the TDWI World Conference in Boston July 20-25, 2014.

About the Author

Fern Halper, Ph.D., is vice president and senior director of TDWI Research for advanced analytics. She is well known in the analytics community, having been published hundreds of times on data mining and information technology over the past 20 years. Halper is also co-author of several Dummies books on cloud computing and big data. She focuses on advanced analytics, including predictive analytics, text and social media analysis, machine-learning, AI, cognitive computing and big data analytics approaches. She has been a partner at industry analyst firm Hurwitz & Associates and a lead data analyst for Bell Labs. Her Ph.D. is from Texas A&M University. You can reach her by email (, on Twitter (, and on LinkedIn (

TDWI Membership

Get immediate access to training discounts, video library, research, and more.

Find the right level of Membership for you.