What is Prescriptive Analytics and Why Do We Need It?


The most innovative companies will stop asking all their workers to become analysts and simply bring the insights to them in the apps they’re using. Prescriptive analytics is how they will make that happen.

For decades, we have been able to use data and analytics to describe what happened. In recent years, we’ve started to use analytics to predict what will happen based on the data we have. This still leaves us with an unanswered question: Now that we have the data, what do we do about it? The answer is prescriptive guidance – analytical insights from data that suggest what actions to take next.

Recent technological advances in automation and AI have made it possible to dynamically

extract insights – not just reporting – from large, disparate data sources and personalize them to every user. Critically, these insights go beyond reporting the present and forecasting the future to recommending actions to achieve the outcomes we want.

This has been the whole point of capturing data from the very beginning. We don’t want to simply know how we did; we want to change what we do next.

See also: The Next Step for Enterprises: Optimization Transformation

Why analytics must become prescriptive

There has always been a gap between raw data and actionable insight. It’s nearly impossible to use data to help decide what to do next when the data is presented as endless spreadsheets or heavily aggregated trendlines. Although we’ve invented tools to help make sense of the data, the first tools were only descriptive, revealing historical information but leaving it to us to interpret what it meant. Recent predictive analytics tools are better but leave us with the same problem.

An additional issue is that these tools are hard to use and often require specialized skill sets and additional degrees. Most business professionals–the very people who want the insights to make smarter decisions–simply don’t have the time to learn these additional skills or add extra steps to their workflows. This creates another gap between the people who can interpret what the data means and the people who can wield the tools to query the data in the first place.

Prescriptive analytics closes both gaps by using AI to automatically analyze data and extract the most relevant insights and suggestions on what to do next. This is ultimately what most knowledge workers want: They want to take action, and the most data-driven action possible. They also would rather not have to learn unrelated skills or wait on IT-led reports.

This doesn’t mean descriptive and predictive analytics are dead. They each have their role in equipping us to make data-driven decisions, and they can often work together. Every company must know what its quarterly numbers are. Executives must always forecast their revenue targets and goals. However, prescriptive analytics is revolutionary rather than evolutionary because their insights are far more accessible to everyone and can be integrated into the operational decisions made continuously throughout the day.  

See also: Augmented Analytics: A New Dimension to Analytics & BI

Prescriptive analytics in action

As consumers, we regularly make decisions and take action based on prescriptive analytics, and often without even knowing it. Music streaming services take our historical listening habits, those of our friends, and even the top charts to suggest new songs to listen to. Mapping apps offer alternative routes due to traffic conditions. Fitness devices suggest a workout exercise so that we meet our daily health goals.

These suggestions are prescriptive analytics at work – the culmination of huge amounts of data across many sources but boiled down to insights that help us decide what to do next. This doesn’t mean we follow these suggestions blindly. Ultimately, we are always the ones to take action based on our instincts and knowledge as people. The insights simply help ensure we make the smartest decision possible.

When we apply the same prescriptive analytics to our business apps, knowledge workers can simply focus on the decision they need to make, often unaware they’re using data or analytics at all.

Customer Service reps will be notified of accounts likely to churn so they can proactively reach out. Sales leaders will not only be able to predict that they will miss quarterly revenue quotas but know exactly which accounts to target to make up for the gap. Retail managers can shuffle and optimize inventory before items sell out by combining sales data, purchasing patterns of other items, external market trends, and even competing promotional campaigns.

There are countless other examples, and for each of them, workers simply have the insights from data they need to take action, exactly when they need them.

The role of embedded analytics

We can’t talk about a future of prescriptive analytics and actionable insight without also discussing how we deliver those insights to knowledge workers. Very few people will take advantage of the smartest insights if they still have to sign in to separate portals or load standalone dashboards just to view those insights.

This is where embedded analytics – or infusing analytics within key workflows – is critical. To truly take advantage of prescriptive analytics so knowledge workers can use insights from data to take action, we have to pair the two together. We have to infuse the insights where we want to take action. Just as data was less useful when we had to wait weeks or months for IT reports to come back, data will continue to be left unused if we have to deviate from our workflows just to dig through dashboards for information on what to do next.

Instead, the most innovative companies will stop asking all their workers to become analysts and simply bring the insights to them in the apps they’re using. Prescriptive analytics is how they will make that happen, using automation and AI to proactively help all knowledge workers take action and make smarter, data-driven decisions.

Scott Castle

About Scott Castle

Scott Castle is an analytics infusion pioneer bringing over 25 years of software development and product management experience to his role as VP & GM Products at Sisense. Scott is passionate about turning data teams into superheroes that find unexpected insights in big data and disrupt traditional BI. Previously, Scott held technology positions at companies including Adobe, Electric Cloud and FileNet. Scott holds computer science degrees from the University of Massachusetts Amherst and UC Irvine. Connect with Scott on LinkedIn, and Twitter.

Leave a Reply

Your email address will not be published. Required fields are marked *