Operational Analytics: Five Tips for Better Decisions

PinIt
operational analytics

To make better decisions using data requires operational analytics. That can include embedding analytics on different systems and using disparate data,

To be useful, analytics must be brought to decision-makers, according to a Sep. 29 TDWI webinar. And that could mean embedding analytics in new areas such as mobile devices and using new kinds of data, such as social media text and geospatial data.

“In the past when people talked about operationalizing or embedding analytics, they were often talking about embedding these things into dashboards or visualizations in their C.R.M,” said Fern Halper, director of TDWI Research for advanced analytics. “But that’s really changing. What’s happening now is they are talking about doing this in interactive dashboards. And they’re embedding analytics in mobile devices” or “databases, or services, or even into systems.”

Analytics can be interactive, integrated (as part of an application) or automated, in which case the aim is to change behavior. No matter what the flavor of analytics, TDWI offered five tips for better decision-making:

1. Make analytics actionable

According to a forthcoming TDWI best practices report, approximately 75 percent of respondents stated that analytics should inform action. That could mean everything from a static report showing sales down in a certain location—leading an executive to take action—to interactive call-center dashboards that determine customer satisfaction is high when a call takes less than three minutes. It can also involve automated analytics such as preventive maintenance or an online sales promotion to a certain group of people. According to a TDWI survey, while 34 percent of respondents receive automated alerts from their analytics, less than 20 percent of respondents actually embedded automated analytics. However, making small decisions over and over again—an area called “decision management”—is a future area of operational analytics, Halper said.

 2. Use Different Sources of Data

“It’s important to consider a range of analytics,” Halper said, including descriptive analytics such as monthly sales reports; predictive analytics; and prescriptive analytics such as scoring likely buyers.

Data sources can be structured, such as clickstreams, sensor, or point-of-sale data; or unstructured, such as social media data, mobile data, satellite images, and video. Approximately 94 percent of respondents to a TDWI survey use structured/transactional data today. More than 44 percent, however, are not using mobile app data now but plan to do so in the next three years. Survey respondents also said in the next three years they plan to use social media text data (42 percent), geospatial data (32 percent) and real-time streaming (39 percent).

“For example, I might be trying to predict customers who churn,” Halper said. ” I might build all sorts of models for this using my structured data from my C.R.M.” or other systems, “then I merge them together with some third-party data like demographic data. But it might be that once I bring in the text data from the call center notes that I really find my predictor.”

3. Think About a Range of Analytics

Analytics can be descriptive, such as last month’s sales; predictive, such as ranking prospective buyers; to prescriptive, which recommends a course of action. “You need the right tool for the right job,” Halper said. “You can do descriptive statistics on data and query it and slice and dice it all day long but there may be a better technique that’s going to help you to get insight.”  According to a forthcoming TDWI report on emerging technologies, predictive and web analytics are becoming more popular. Fifty-seven percent of respondents to a survey use predictive analytics today and another 37 percent plan to use it in the next three years.

4. Consider in-memory technologies

Typical business intelligence relies on database tables managed on disk. But as memory costs fall, accessing RAM is much faster. “In-memory offers data access that “can be orders of magnitude faster,” Halper said, and significantly increases application performance by reducing data latency bottlenecks.

5. Develop an innovation strategy

TDWI advises to tie analytics to revenue goals and consider the organizational and data infrastructure.

“If you know what analytics you want to use but don’t have the data infrastructure to support it or the skill set to analyze it, then your plan goes nowhere,” Halper said.

She advised that the plan should start “with the vision of where you’re going to be a few years from now. Have a vision. And it’s important to tie the plan to something important like revenue. That gives the project a better chance of success.”


Want more? Check out our most-read content:

Frontiers in Artificial Intelligence for the IoT: White Paper
IoT’s Value? It’s Mostly in Traditional Business
Research from Gartner: Real-Time Analytics with the Internet of Things
Report: Text Analytics Must Move Beyond Social
Why Enterprises Struggle with Apache Storm and Spark Streaming

Liked this article? Share it with your colleagues using the links below!

Chris Raphael

About Chris Raphael

Chris Raphael (full bio) covers fast data technologies and business use cases for real-time analytics. Follow him on Twitter at raphaelc44.

Leave a Reply

[Malvinas]
[Malvinas]
[Malvinas]
[Malvinas]