Predictive analytics is leveraging historical data and business trends to anticipate the probability of certain scenarios occurring, ideally helping to estimate the likelihood of a future outcome based on historical data patterns.
There are actually four types of analytics starting with Descriptive, Diagnostic, Predictive and Prescriptive. In simplest terms, descriptive analytics is “what happened”, diagnostic analytics is “why did it happen”, predictive analytics is “what will happen” and prescriptive analytics (actionable intelligence) is “what should I do”. Technically all four types analyze large volumes of data to identify business trends and “events” that could impact business decisions.
However, predictive analytics requires users or workers to understand and know how to interpret “the future”. Whereas, actionable intelligence is analytics for everyone (including those at the edge) with a focus on future performance through identifying controllable factors and providing actionable opportunities that deliver results.
What Is Actionable Intelligence?
Actionable intelligence is one of the most advanced forms of business analytics. It uses machine learning and pattern detection rules and algorithms to identify anomalies in a company’s operations – and then prescribes a corrective action to optimize the outcome. That last part is what makes actionable intelligence so valuable: it can intelligently prescribe the action for someone to take to optimize for a particular business outcome.
It can analyze a company’s data and tell them:
- What is happening?
- Why it happened? (root cause)
- Who should respond?
- How they should respond in order to positively impact revenue
Learn more about Actionable Intelligence in our dedicated What Is Actionable Intelligence FAQ.
What Are the Key Differences?
Predictive Analytics is an analytics model which takes into account patterns in historical data and historical timelines to inform you and predict what will happen next, whereas Actionable Intelligence simulates different actions that you could take based on a goal that you want to achieve (e.g. increase sales, reduce shrink) to discover the most optimized outcome.
What Are the Examples of Predictive Analytics and Actionable Intelligence
Real-life examples of how predictive and actionable intelligence can be used are:
Preserving Product Quality
A clothing retailer used actionable intelligence and pragmatic AI to identify and resolve a unique return opportunity. A new product was experiencing a high rate of returns across the majority of stores in the initial days of sales, plus a high rate of damaged product. This opportunity was sent to the merchant with the recommendations to (1) evaluate the product quality, (2) check the label, and (3) contact the vendor for an allowance increase or replacement product. It turned out that the product was missing the correct washing instructions on the care label; the actual instructions on the label would damage the item. As a result, an action was created for the stores to attach an updated garment-care label to the remaining affected merchandise. A similar notification was sent to the manufacturer with the correct care instructions. Ultimately, product returns fell 78 percent in the weeks following corrective action and these actions helped save the style.
A retailer used actionable intelligence to identify a significant spike in large-egg damages in a subset of stores within a specific region. The merchandising department contacted the vendor to find out why this was happening. It turned out that the vendor had suffered a fire at the plant that produced the large-egg cartons, so they began using medium ones to ship out the large eggs. These cartons were too small, causing additional friction that damaged the large eggs. Without identifying and correcting the anomaly, the grocer may never have realized the excess damages, nor would it have been able to trace the root cause and secure a credit from its vendor.
Match Customer Profiles
A fashion retailer used actionable intelligence to successfully reduce aged inventory and increase sales by offering near end-of-season items to households that matched the style and age range of a likely buyer based on loyal household buying patterns. For example, a targeted promotion could be sent for a white skirt offered to a 25-30 year old female at a 30% discount, enticing her to come into the store to take action now and she’ll be able to wear the skirt through the end of the summer. Without actionable intelligence, a lot of retailers wait and then discount the item at 75% off for everyone and hope that same female shopper comes in, but now they’ve lost profits.
Buy with Demand in Mind
A prescriptive-action pattern identified a specific product that had a higher Cost to Serve (cost to get the product to the customer within the promised timeframe). The opportunity was sent to the DC to resolve. They quickly noted that the products in question were purchased in a unit of measurement and with packaging that did not align with the customer demand or DC processing. The action taken was to notify the buyer to create orders specific to eCommerce. This improved the DC’s fulfillment speed. Communication across channels improves performance!
Increase Basket Size Through ‘Buy-With’ Pattern Identification
A retailer used pragmatic AI with actionable intelligence to identify that when specific shoppers purchased broccoli they also purchase bananas. This finding was communicated to the enterprise along with prescriptive actions to capitalize on this finding. Store managers were instructed to reorganize displays so the bananas and broccoli were the appropriate distance apart to increase baskets that included bananas and to also help drive up impulse buys.
Reward Loyalty, Not Fraudulent Activity
A large retailer with a popular loyalty card program used actionable intelligence to uncover an issue they weren’t even aware was a problem. Leveraging analytics, they were able to discover a trend that shoppers were cheating their reward points for free rewards. Their retailer’s previous system did not monitor points scanning until the end of each day, so if a customer brought their receipt to multiple stores in a single day, it was easy to receive many times more reward points than were due. The actionable intelligence solution caught the recurrence and recommended real-time tracking to ensure receipts could only be scanned once for points, saving as much as $100 per shopper, means millions to the bottom line.
Increase Items Per Transaction - Why Just Buy One?
Actionable intelligence was used by a retailer to identify stores with a high rate of single-item transactions, as well as the associates that were driving this behavior. An opportunity was sent to the stores to reschedule these associates with associates who regularly rang multiple items per transaction, so associates would learn how to upsell. The results were an immediate uptick in sales, items per transaction, and peer-peer training. A big win-win for the organization.
Increase Brand Loyalty
A grocer had a consistent shopper base that made one large shopping trip ($120+) and one or more smaller trips per week. When a competitor moved in close by, the grocer experienced a drop in the smaller, supplemental trips. With actionable intelligence, the grocer identified the households that had stopped making the smaller trips, as well as the particular items they were buying elsewhere (i.e. meat or vegetables). With this information, the retailer was able to send targeted offers to those households, encouraging them to return. Ultimately the retailer won back many of those customers, leading to increased profits.
These are just a few examples of how retail leaders can use practical AI to look critically at business operations and take action to realize returns. Whether there are delays in shipping, customer complaints on product durability, or a seemingly isolated incident of broken merchandise, practical AI and actionable intelligence go beyond complicated tables of data to give actionable insights. Through that lens, pragmatic AI is the natural next step for analytics adoption.