Predictive analytics isn’t just for oil and gas exploration anymore. The power of predictive analytics is being injected into a wide range of revenue-focused initiatives across all industries.
In 2018, a third of businesses in the EIU’s Intelligent Economies study said predictive analytics was already the most frequently used AI technology in their organization. Almost two thirds of the CIOs in Capgemini’s most recent World Quality Report said they’d be focusing on predictive analytics in 2019, but what differentiates predictive analytics and where can you get value from it in your organization?
Predictive analytics differs from business intelligence primarily in perspective: whether you’re looking forward or backwards with data. With BI, the emphasis is on reporting and visualization — slicing historical data to understand what has happened. But with predictive analytics, “you’re no longer talking about descriptive analytics and you’re mostly focused on building a model for predictions,” says Kjell Carlsson, a senior analyst at Forrester.
Many of those algorithms are also used for machine learning, and Carlsson views predictive analytics and machine learning technologies as complementary. But predictive analytics doesn’t have to be complex. Salesforce Einstein Discovery and the insights feature in Microsoft’s Power BI both use regression analysis, but because they can work on massive data sets, they can find insights that would be too tedious for business users to discover on their own.
“If I’ve got a solution that guides salespeople to focus on the accounts that have the highest likelihood to convert and gives them reasons why this is a good account to reach out to right now, like they just downloaded a white paper, then that becomes extremely valuable from a business point of view,” Carlsson points out.
Predictive analytics is likely already in use at your organization, driven by lines of business rather than IT, he warns. “There is an incredible amount of shadow IT here,” Carlsson says. That can be a problem if poor data governance leads to a data breach, but there can also be issues when successful prototypes need wider deployment and longer-term maintenance.
At that point, CIOs and chief enterprise architects are being asked to take over. To keep ahead of the game, here are seven key projects primed for use of predictive analytics today.
1. Predictive equipment maintenance
Knowing when industrial or manufacturing equipment is likely to break down can help save money and improve customer satisfaction. Elevator manufacturers, air conditioning systems, national railways and oil well operators use IoT sensors and digital twins to provide predictive, proactive maintenance.
Here, predictive analytics doesn’t just help you avoid outages and repair bills. Knowing which spare parts, equipment and trained staff will be needed means work can be planned more efficiently, with fewer trips to the site and no delays waiting for the right part. Plus, it’s faster to repair a part before it fails because there can be damage caused by the failure. Avoiding that also extends the life of the machinery.
The information you collect can also feed forward into product design for the next version or help you develop better operating procedures.
2. Predictive IT
Predictive maintenance is also a boon for IT. Data center management tools, such as Nlyte or Virtual Power Systems, can warn you to replace UPS batteries or perform maintenance on a cooling unit.
“If you buy storage from Dell, the ProSupport Plus service uses predive analytics to predict when drives are going to fail and they pre-emptively send you replacement drives before they fail rather than afterwards,” Carlsson says. Similarly, Veritas offers Predictive Insights for it storage appliances that creates system reliability scores. When those drop, the IT team might get a notification to install a patch — or Veritas might send out a technician to replace a part before it fails. HPE’s 3PAR InfoSight management and DataDirect’s Tintri Analytics use predictive analytics to improve storage performance and handle routine storage management.
This is one area where a third-party service is likely better than building your own because if you may not have enough data to predict problems, Carlsson points out. “External vendors have the advantage of collecting data from different customers. If there’s an update for your particular hardware that’s causing problems for other companies with the same configuration as you and you haven’t applied that patch yet, you will never know from your internal data that there’s anomalous behavior.”
Predictive IT doesn’t have to be hardware either. Windows Server 2019 has predictive analytics built into the Windows Admin Center to help you perform capacity planning for compute, networking and storage, including clusters. System Insights uses local data such as performance counters and system events, and you can write your own predictive maintenance capabilities for performance, say, and then use Azure Monitor or System Center Operations Manager to view predictions across groups of servers.
3. Forecasting HVAC needs
Combine the weather forecast with what your building automation system tells you about how your facilities are used by staff and the data you can get from your HVAC system and you can reduce costs for heating, ventilation and air conditioning.
It takes time to get a building to the temperature you want when people are at work (especially if you’re saving energy by not heating or cooling them out of hours), and that varies for each building and depends on the weather. Plus, not every building is fully occupied all year round. Instead of starting the systems at the same time every day for every building, you can save money and keep employees more comfortable at work by predicting the right time to ramp up the HVAC system. When Microsoft’s real estate team applied this to just three buildings, they saw savings of $15,000 annually; that will turn into more than $500,000 once the system is in 43 buildings — and 60 fewer hours when employees are sweating or shivering.
4. Customer service and support
Predictive analytics is common in sales tools like Salesforce, but you can also use it to handle the customers you already have, whether that’s field service or call centers. Adobe Analytics uses predictive analytics to forecast future customer behavior down to when you’ll run into special shipping requirements.
MTD makes outdoor equipment like lawn mowers and snow ploughs and credits the predictive analytics and real-time information it’s added to call center systems with reducing call abandonment by 65 percent and cutting the average time to handle a call by 40 percent, thanks to better agent scheduling — because managers know in advance when they’ll need more agents at work.
Ecommerce sites have long had the advantage of being able to track customer behavior to help predict sales figures. Jet.com even models how likely it is that a supplier will have the right amount of inventory in stock before listing products in its marketplace. Now retail stores are turning to IoT sensors and predictive analytics to forecast what, when and where customers will buy, to help with inventory management. Polo and Urban Outfitters are using shelf-counted cameras and Trax’s predictive analytics system (running on Google Cloud) to do real-time stock tracking and management.
Dr Martens is using a mix of IoT, predictive analytics, machine learning and Dynamics 365 to understand more about the demographics and buying patterns of the customers who are browsing their stores. Sales staff can then use this information to make suggestions or even rearrange where products are displayed using custom schematics for the store.
6. Quality Assurance
Predictive analytics is ideal for QA, because whether it’s testing physical products or part of DevOps, QA is about avoiding defects, problems and mistakes by assuring risk. You can determine patterns and predict potential risks based on trends and use predictive analytics to reduce cycle times and cost by targeting testing where defects are most likely to occur, says Darren Coupland, Deputy CEO and COO at Sogeti UK (part of Capgemini).
“CIOs should be using predictive analytics, along with AI and cognitive solutions, to truly understand the quality of their overall operation and make informed decisions based upon insights. In order to take this one step further, CIOs should consider combining additional data sources, such as PPM [project portfolio management] tools, SCM [source code management] tools and operational tools, in order to predict the successful delivery of projects and provide important information into the overall business risk associated with a change,” Coupland says.
7. Alongside business intelligence
If you want to give business teams the freedom to work with predictive analytics alongside the more familiar visualization and analytics tools, and still have central oversight, the new no-code AI tools that will be in public preview for Microsoft Power BI soon may be what you’re looking for.
Power BI has been able to do simple predictive analytics like forecasting future patterns for time series data, with sliders for the confidence level and how strong you expect seasonal trends to be. You currently need to build more sophisticated models in a tool such as Azure Machine Learning Studio and use R scripts to extract data from SQL Azure and send it to the machine learning model and then extract the scores into Power BI. With the new no-code connection, business analysts will be able to choose and train a model in Azure Machine Learning Studio and apply it to Power BI data without leaving the Power BI interface. Your data science team can also create and train models with the Azure machine learning tools for them to use that will show up in in Power BI automatically if a business user has access to them.