Predictive or Prescriptive Analytics? Your Business Needs Both

Predictive vs. prescriptive analytics

Predictive and prescriptive analytics inform your business strategies based on collected data. Predictive analytics forecasts potential future outcomes, while prescriptive analytics helps you draw specific recommendations.

Predictive and prescriptive analytics are tools for turning descriptive metrics into insights and decisions. But you shouldn’t rely on one or the other; when used together, both analytics types can help you shift your business strategy to create the best possible outcomes. 

“Predictive by itself is not enough to keep up with the increasingly competitive landscape,” said Mick Hollison, president of enterprise data management company Cloudera. “Prescriptive analytics provides intelligent recommendations for the optimal next steps for almost any application or business process to drive desired outcomes or accelerate results.”

What is predictive analytics?

Predictive analytics is an advanced analytics category that helps companies make sense of potential outcomes or a decision’s repercussions. By leveraging mined data, historical figures and statistics, predictive analytics uses raw, up-to-date data to peer into a future scenario.

Until a few years ago, predictive analytics was the province of enterprise-level businesses – the only ones able to afford to parse and interpret reams of data from multiple sources. However, the growth in software as a service (SaaS) providers and CRM analytics means even small companies can access valuable data analytics. 

A key aspect of predictive analytics involves segregating superfluous or misleading data that could distort the insights. For example, a travel company with a sales rep in every state shouldn’t emphasize data provided by an employee in Alaska. 

What is prescriptive analytics?

Prescriptive analytics also looks at future scenarios, but it employs a more technological approach. It uses complicated mathematical algorithms, artificial intelligence and machine learning to take a deeper look into the “what” and “why” of a potential future outcome. 

Prescriptive analytics can also help a company see multiple options and potential outcomes. As more data comes in, prescriptive analytics can alter its predictions and suggestions accordingly.

“Prescriptive analytics can help companies alter the future,” said data-driven digital strategist Immanuel Lee. Predictive and prescriptive analytics are “both necessary to improve decision-making and business outcomes,” he added.

Examples of predictive and prescriptive analytics in action

We use predictive and prescriptive analytics in our everyday lives. Here are three examples of predictive and prescriptive analytics working together.

Navigation apps

Motorists rely on GPS-enabled navigation apps to get from point A to point B. GPS navigation is equally essential for small businesses that rely on delivery services. Predictive analytics can take existing GPS-sourced travel data and map a potentially faster route. 

Thomas Mathew, chief product officer at influencer engagement platform Zoomph, said that’s where the effort starts. “Prescriptive analytics builds on [predictive analytics] by informing decision-makers about different decision choices with their anticipated impact on specific key performance indicators.” 

For example, consider the traffic navigation app Waze, which blends multiple factors to respond to users’ origin and destination input. The app advises you on different route choices, each with a predicted ETA. “This is everyday prescriptive analytics at work,” Mathew said.

Inventory planning

Retailers need to know how much stock to order to fill their shelves. While many retailers rely on educated guesses, analytics can help them plan a more precise inventory management strategy.

Guy Yehiav, president of SmartSense by Digi, said that as the retail landscape changes, businesses can use prescriptive analytics to clarify predictive data and improve their sales plan. 

Yehiav gave the example of a retailer offering free expedited shipping to loyal customers. Based on past customer behavior, a predictive model would assume that customers will keep most of what they purchase with this promotion. However, imagine a scenario where one customer purchases eight items of clothing before returning all but one.

“The retailer paid for expedited shipping with the assumption that there’s this great consumer out there who bought eight items, so they’re willing to invest and lose a little margin” on shipping, Yehiav said. “The algorithm didn’t take [return] behavior into account.”

For this retailer, reducing losses on outlier customers who don’t follow what predictive analytics forecasted means having policies in place to cover itself. Using prescriptive analytics, Yehiav said the retailer might decide to give an in-store-only coupon to customers who make returns (to encourage another purchase in which shipping isn’t a factor) or notify customers that they must pay for return shipping.

Weather forecasts

Predicting the weather can be a dicey proposition, but with the change of seasons comes the shift from indoor activities to fun in the sun. Sporting goods stores comprise one small business sector that benefits from nicer weather and increased physical activity.

If the store’s forecasts indicate that sales of running shoes will increase as warmer weather approaches in the spring, it might seem logical to ramp up the running-shoe inventory at every store. However, in reality, the sales spike likely won’t happen at every store across the country at once. Instead, it will creep gradually from south to north based on weather patterns.

Arijit Sengupta, former CEO of automated business analytics company BeyondCore and founder of Aible, said predictive and prescriptive analytics could help you plan for this scenario.

“To flip the switch on massive running-shoe distribution nationwide would be a huge mistake, even though the predictive analytics indicate sales will be up,” he added. “But with prescriptive analytics, you can pull in third-party sources, like weather and climate data, to get a better recommendation of the best course of action.”

Putting analytics to work

Here are a few tips to help you get the most out of your analytics programs.

1. Start small with data analytics.

Data analytics is a complex subject that can be overwhelming, and you don’t want your best insights to get lost. Lee advised thinking big with your overarching analytics strategy but starting small tactically.

“With the complexity of big data and the systems that manage and process data, we can easily overlook the fact that sometimes there’s a solution in the simplest thing,” he said. “Small wins will help earn support for long-term analytics projects.”

2. Create rich data sets.

There are many what-if scenarios when you run and market a business, and predictive analytics doesn’t always account for alternate possibilities. Mathew said looking at your predictive analytics more closely to create richer information sets (by accounting for demographics like gender and age) will yield better results from your prescriptive recommendations.

“Social media marketers care about maximizing engagement and reach on their social posts,” he said. “Prescriptive analytics can make data-driven recommendations, such as use of a specific hashtag or emoji, to maximize social traction with a specific audience segment.”

3. Understand the reasons behind prescriptive recommendations.

Sengupta emphasized the importance of fully understanding the logic, nuances and circumstances behind the results of prescriptive analysis before taking action. Be prepared to prove that your results are statistically sound.

“Pretty graphs can be very compelling, but this is only software, and its analytical power is only as accurate as the human who designed it and the data we feed it,” Sengupta said. “It’s critical that business users understand the ‘story’ behind the results and the prescriptive action suggested.”

4. Keep your systems up to date.

As your business grows and evolves, so should your algorithms. Hollison noted that both predictive and prescriptive analytics should be updated continuously with the latest data to improve predicted and prescribed actions based on real-time successes and failures.

“Predictive and prescriptive analytics depend on a solid data foundation,” Mathew added. “The analytics are only as good as the data that feed them.”

Analytics tools for small businesses

There’s a common misconception that the analytics industry is dominated by tech giants like Microsoft and IBM, which offer analytics software through their Power BI and Cognos with Watson platforms. Almost inevitably, web services titan Amazon also has a presence in the market with its cloud-based QuickSight BI service.

Alongside these tech giants, there are numerous more specialized entrants into this congested marketplace, including SAP, Zoho and Sigma. Many offer free trials, though some will only reveal costs when you register for a quote. That’s not a step companies tentatively considering analytics are always happy to take.

Some analytics platforms are code-specific (for example, Dash focuses on Python), while others are geared around no-code for simplicity, such as GoodData. Tableau works on Windows or Linux, whereas Domo is cloud-native.

Factors to consider when looking for small business analytics tools include how many data streams you have and what format they come in, how you want to visualize parsed data, and what objectives you’re looking to achieve. For example, InsightSquared studies every revenue activity throughout a business before calculating successful deal profiles and determining where revenue-raising improvements can be made.

Neil Cumins and Andrew Martins contributed to the writing and reporting in this article. Source interviews were conducted for a previous version of this article.