Perhaps it’s better to start off with what SAP’s Predictive Analysis software cannot do: It won’t predict the future for your company’s stock price, whether a freak blizzard will derail next Halloween’s all-important sales, or that a social media slip-up by your VP of marketing will result in the loss of your 100 best customers.
Charles Gadalla, director of advanced analytics at SAP, jokes that when he makes presentations on predictive analysis, he will invariably get those types of questions from the audience. “They expect you to do a fortune telling, or they’ll say, ‘Tell us what are the lottery numbers tomorrow,’” he says.
A crystal ball it ain’t. So just what does SAP Predictive Analysis actually do?
According to SAP, the software can help employees “visualize, discover, and share untapped opportunities and hidden risks—all in real time.” It synchs with SAP’s BusinessObjects suite and, in particular, Lumira, the visualization tool formerly known as Visual Intelligence (read ASUG News’s bumpy test drive of Lumira).
Let’s go just a little deeper: predictive analysis is based on a predictive model, which is “an equation, algorithm or set of rules used to predict an outcome based on input data,” writes Hillary Bliss, the analytics practice lead for BI consultancy Decision First Technologies and author of a new report on SAPexperts.com.
“Predictive models can simply be a set of business rules based on past observations,” she continues in the report, “but to develop more accurate and statistically rigorous predictions, statistical algorithms for developing the predictive equation are often used.”
For SAP Predictive Analysis, Gadalla says a combination of three key components are key: descriptive, prescriptive, and optimization.
Descriptive describes what has happened in an organization in the past, looking at descriptive statistics (mean, mode, median, etc.) that are used to quantify and measure historical events. Prescriptive is the part that analyzes trends and relationships in past events and uses them to make predictions on what might happen in the future. And optimization says: Given a range of outcomes, which are most likely and how can a company best prepare for them?
The last part is critical, Gadalla contends. When thinking about how most companies typically view analytics (hello Microsoft Excel!), the move from an employee asking the data “What happened?” to now tapping into all different types of critical data sources and asking, “What is the best that could happen?” is a huge improvement.
Who Will Use Predictive Analysis?
So (if you’re still with me), who’s Predictive Analysis for—solely those Ph.D. quants who live and breathe statistics, or for the Excel Warrior who can PowerPivot with the best of them?
Gadalla says SAP is going after the business analyst and sophisticated business user audiences (though the data scientists can certainly use the product).
“You can bring in data, massage it, visualize it, and come to something interesting, and then build a forecast model or build a decision tree,” he says. What’s new is that SAP customers running, say, CRM or SCM can, with a “couple of clicks of a button,” use Predictive Analysis to tap into those specific data pockets as well, he adds.
The “Before Predictive Analysis” processes he describes center on business users either waiting for specialized PhD types to do the heavy lifting on scripts, or they could rely on good ‘ole Excel and make an educated guess. “We’re trying to plug that hole,” Gadalla says, “and give [business analysts] more sophistication.”
Decision First’s Bliss says someone who’s used to working in Excel or has used Webi or Explorer will be able to use Predictive Analysis without too much effort.
“It’s certainly easy enough to use,” she says. “This doesn’t require learning a whole new programming language. It’s all point and click. And the visualizations are very intuitive, with a lot of drag and drop functionality.”
What’s New in Version 1.0.11
SAP announced Service Pack 11 in late June, touting tighter integration with Lumira for model visualization, expanded flexibility in connecting with other SAP applications and faster execution against large data volumes, among other enhancements.
Bliss points out that version 11 offers the ability to code your own R functions within the tool. “So that really opens up for the quants and statistics people to use the tool to access the visualizations, as well as the ability to share those resources with Lumira,” she says.
That means the more sophisticated users can package up the algorithms they’re building and share them with those (mere mortal) users who can’t code in R. The result: Reduced silos and a special kind of sharing that would make a kindergarten teacher smile.
Bliss talks excitedly about a test environment she’s been playing in that uses the data from an online retailer and then combines it with BusinessObjects universes, reports and dashboards, SAP HANA and Predictive Analysis.
To the lay person (like me), the possibilities seem endless in what, say, a retailer could do “to generate models to better understand customer behavior and value,” she says. The work, she adds, is an example “of how we can use existing BI infrastructure and really integrate Predictive Analysis right into an existing environment and leverage BOBJ tools to create an implemented solution in a matter of a few hours.”
Companies such as eBay are going down the Predictive Analysis-plus-HANA path. “SAP HANA is able to analyze 500 metrics [at eBay] and select the best model automatically to determine true positives with 100 percent accuracy at 97 percent confidence,” notes an SAP.info profile of eBay. “The reliable automated detection system will free up bandwidth so eBay analysts can spend their time on strategic tasks and eBay can focus on what it does best: connecting buyers with things they need and love.”
Buying Predictive Analysis
So how does one actually buy Predictive Analysis? There are, it turns out, several different methods, and I’ll do my best to make it clear:
1. It’s bundled into SAP Lumira, within the Lumira code base, Gadalla says. But there’s a catch: You have to buy a key code to “light it up,” he says.
2. But if you buy Predictive Analysis, you get Lumira for free. (Got it?)
3. Predictive Analysis is embedded within about 20 different HANA apps (such as CRM, SCM, Partner Relationship Management and Liquidity Forecaster).
4. Predictive Analysis is also bundled with a HANA box, “where I’m using HANA as a server, where I’m doing my analysis,” Gadalla says.
Current pricing on SAP Predictive Analysis is $20,000 per seat with a minimum purchase of 5 seats, according to Gadalla.
No discussion of a new SAP product would be complete without that requisite HANA mention, of course. Gadalla reports that HANA went head to head with top competitors in doing clustering analysis, and he claims that competitors’ wares took 40 hours to do the job whereas HANA took just 2 minutes. “It was the same exact equation and same exact data,” he says.
But whereas HANA can giveth, it can also taketh away in that some SAP customers (and pure BOBJ shops) are still reluctant to embrace HANA. “It’s a double-edged sword: There are people who love it, and there are those who are leery,” Gadalla says. “We are tying our carts firmly to the HANA horse, because without HANA, [SAP Predictive Analysis] is just another mathematical tool.”
Decision First’s Bliss says that with the massive amounts of corporate data today, “you really have to have that capacity to run the larger data sets [on HANA]. And I think users are going to expect that in a tool.”
As to how many Predictive Analysis customers there are today, Gadalla won’t give a precise number, only to say that number is in “the triple digits.” According to a recent study by SAP Performance Benchmarking, nearly 70% of companies surveyed reported limited or even no capabilities to perform predictive modeling with big data.
Clearly there’s much more work on “predictive analytics awareness” to do to increase that number of customers. Bliss says that the education piece surrounding predictive analytics will be key going forward for SAP, saying: “It’s about making users aware of what’s actually available.”