Descriptive, Predictive, and Prescriptive Analytics: What’s the Difference?

February 10, 2020

Descriptive, predictive, and prescriptive analytics use data to answer fundamental questions that help guide organizational strategy. If you have any interaction with data at all, you’re using one of these analytics styles to glean insights. But because data analysis is still new in many professional environments, there’s plenty of confusion around these three terms.

How do you know which type you’re using? Which type you need? And how do you know if it’s the right time to implement a new analytics model? Ultimately, it’s not about choosing the right one, but learning how all three work together.

Descriptive | What happened?

If you’ve had limited experience with data, descriptive is likely the kind you’re familiar with. Using descriptive analytics, organizations can generate crystal clear hindsight from mined data. Rather than launching projects and being uncertain of the full outcome, descriptive data tells a detailed story of what happened (e.g. How many units were sold and when? How many people signed up for the mailing list last month, and where were they from?)

Descriptive analytics is useful in many circumstances, including:

  • Comparing figures from different time periods
  • Assessing the success of a marketing campaign
  • Generating reports that show progress over time
  • Revealing problem areas where goals are not being met

With simple data visualizations, descriptive analytics allows organizations to get a full picture of what happened: whether last week, last month, or last year. With real-time analytics, many organizations even have access to what happened 5 minutes ago. While descriptive may be seen as ‘old news’ in comparison to the next 2 analytics models, it’s still highly relevant and valuable today.

At its core, having descriptive data is like having a mathematical genius on hand who can spit out numbers, facts, and figures about your organization in a moment’s notice.

Predictive | What will happen?

Predictive analytics goes a step further, taking past data and using it to forecast what will happen in the future. This is done using a machine learning model that assesses trends and patterns.

For example, predictive analytics is useful for:

  • Forecasting how much inventory will be needed
  • Predicting what product features customers will respond to
  • Automating follow-up emails with customer segments

This doesn’t mean that predictive analytics can tell you what will happen with absolute certainty. It simply uses all the information available to generate a statistical best guess, and often this is much more accurate than human assumption. With predictive analytics in place, organizations are better prepared to make quick decisions and avoid being surprised by a mishap.

A perfect example is Starbucks use of IoT devices that predict when a piece of machinery is likely to need service. With several data points, Starbucks can prevent last-minute repair costs and potentially disgruntled customers (for more on this, check out our blog: How Starbucks Employed AI for a Better Customer Experience).

“Research firm Aberdeen found that companies homing in on customer needs and wants through predictive analytics increased their organic revenue by 21% year-on-year, compared to an industry average of 12%.” source

At its core, predictive tools are akin to having a genie in your office who seems to sense things and provide ample warning before they happen.

Prescriptive | What should you do about it?

Running a business has long been considered a risky endeavor, and startup failure rates remind us of this all the time. Not knowing what actions to take – especially when things aren’t going well – is a struggle even the largest, most successful companies know well.

Prescriptive analytics can help you:

  • Determine what factors are impacting costs
  • Identify what changes may improve an employee training program
  • Assess which of two decisions may yield a financially safer outcome

Predictive analytics goes a step beyond both descriptive and predictive analytics and provides actionable advice: Considering what this data shows, what is the next step to take? Prescriptive analysis shows some possible decisions you can make and what their implications are likely to be, giving you the best possible chance of making an informed decision.

Google’s self-driving cars are one of the most famous examples of prescriptive analytics and AI being used in groundbreaking ways.

At its core, having prescriptive technology in your corner is akin to having an extremely wise business strategy consultant on hand 24-7.

In Tandem

By now, you can start to see how these analytical components are not competing – nor do you need to frantically ditch your old strategies the minute a new analytics tool is released.

A strategy that encompasses multiple data analysis methods can help you build a beautifully flowing system that runs your organization like clockwork.

Advanced analytics operates under rule-based algorithms, but it is anything but fixed. It can help you initiate radical changes and pivot your strategy as often as you need to. This iterative process allows you to refine your data usage continuously over time.

In the near future, organizations will graduate from putting out fires and scrambling to keep up with customer desires. Instead, the data revolution that’s currently unfolding will allow for an unprecedented level of confidence and calm. Rather than clinging to rigid operational models, organizations can continuously adapt, learn, grow, and take new action. A data-driven strategy is one of constant optimization, and thus a constantly evolving ROI.

Looking for an easier way to put your data to work for you? Get in touch with us to start building your systems of engagement, intelligence, and optimization.

If you liked what you’ve just read, we have more…