Descriptive, Predictive & Prescriptive Analytics. Which Data Analysis Model Do You Need?

analytics.jpg

Data-driven applications and services are becoming more mainstream as the cloud revolutionizes the data analytics scene. Among different data scientist circles, there is a lot of discussion on the evolution in the way analytics are leveraged.

Over the past several years, Hadoop, NoSQL, and major cloud providers created new avenues for big data storage and analysis. I recently attended a webinar with a commissioned report from 451 Research that illustrated the evolution of data analytics from user-driven to automated workflows, using the robust capabilities of Amazon Web Services.

the big data analytics landscape
the big data analytics landscape

(451 Research Report)

What used to take four to five hours in an on-premise data warehouse can now be completed in 10-12 seconds with AWS. Analyzing big data sets requires a significant amount of compute and storage capacity. A Platform-as-a-Service (PaaS), such as AWS, provides an ideal model for data analysis. You can scale up or down based on requirements and only pay for what you use. Without having to wait for hardware, you can quickly change your compute and storage resources to suit your exact needs.

So, what are the possibilities of a big data analytics suite in AWS? Here are some of the most common use-cases and types of big data analytics in the cloud.

Descriptive Analytics

Descriptive analytics is often considered the first start in data analytics. Descriptive analytics takes massive amounts of historical data and turns it into digestible chunks. It is the process of deciphering what happened in the past and turning it into something interpretable for the past. It can be helpful to understand past behaviors and consider future outcomes. Most descriptive analytics falls in line with statistical modeling.

Here are some common business questions answered using descriptive analytics:

  • How much did this customer segment spend during a given period?
  • What promotions did this customer segment engage with?
  • When did this customer segment spend the most?
  • What is the value of this customer segment based on past spend?
  • What behaviors did this customer segment take during a given promotional period?

Predictive Analytics

Businesses can leverage predictive analytics by using statistics, computational modeling and machine learning to identify new sources of data and competitive insights. Using historical data as a guideline with a known set of outcomes, patterns can be found to predict future actions with a surprising degree of accuracy. It enables the business to forecast on what may happen in the future based on probabilities. Think of this data type as information that can be rolled back into multiple channels for the business. We have seen it leveraged in financial modeling, supply chain, logistics and transportation, productivity monitoring, order processing, and more. It offers the business a means to provide actionable insights with your data.

Here are some common business questions answered using predictive analytics:

  • How much revenue will my business generate during the next peak promotion?
  • What is the revenue potential of my customers in a city or region?
  • Are customers most likely interested in Product X or Product Y?
  • What is the probability that these customer sets will visit our store?
  • What are additional products that this customer finds interesting?

Prescriptive Analytics

Prescriptive analytics goes even further than descriptive and predictive analytics by allowing a data analyst or scientist to “prescribe” on possible outcomes through data insights, such as past business outcomes, new algorithms, and advanced statistical modeling. You can leverage data and prescriptive analytics to predict trends, future behavior patterns, and finally provide business decision trees to take full advantage of your data insights.

Here are some common business questions answered using prescriptive analytics:

  • What’s the next promotion that I can offer to this customer segment?
  • How do we optimize production on these widgets?
  • What is the best course of action to optimize our customer journey?
  • Which product is the best fit for our customers?

With any of the analytics types above, you can understand your business at various levels and make more informed decisions for future outcomes. Less than 3% of companies are using prescriptive analytics in their business, and only 30% are using predictive according to recent Gartner big data reporting. Digital transformation and competitive advantage rest on the untapped resources of data analytics. For more information on building a use case for the analytics types above, please reach out to us for a customized consultation for your business needs.

Always be in the Know, Subscribe to the Relus Cloud Blog!