In the early days of 2014, a handful of analysts said predictive business analytics would start to catch on within the office of finance this year. But three months in, the use of predictive analytics is still the exception rather than the rule, according to experts.
Why the low adoption? Experts say several impediments stand in the way of widespread use of predictive analytics -- some technology-oriented and others having to do with people and processes.
One roadblock, according to Lawrence Maisel, founding partner of DecisionVu and author of the book Predictive Business Analytics: Forward Looking Capabilities to Improve Business Performance is simply a resistant or lackadaisical attitude toward change.
He used the analogy of quitting smoking. While smokers know the habit is bad for their health, and they'd like to stop, the days just keep rolling along with cigarettes in hand. But if they're really serious about quitting, one day they'll have to set the plan in motion.
"It's the notion of seeing the value of the change. You're only going to change once you really understand the value proposition" and internalize a commitment to a new way of doing things, Maisel said.
And the value proposition isn't only about the benefits of predictive business analytics -- it's also about keeping up with competitors. Companies not taking steps toward implementing predictive analytics today could be setting themselves up for future failure.
"There's a consequence to not adopting these technologies and so far that's been measured in terms of small drops in profitability and customer retention," Maisel said. "[But] the real key is you're not growing. It's the wrong slope of the curve, as they say."
Time, skills and disjointed data block predictive analytics adoption
One of the major benefits of predictive analytics is being able to spot and respond to business trends faster, Maisel said.
"If you see a trend beginning to emerge, you can then take actions to change that path," he said. "If they're positive events or results, you want to emphasize and leverage [them]. If they are adverse events, you want to quickly take corrective action."
Steve Meek, president and CEO of the Fulcrum Group Inc., a technology consulting firm based in Keller, Texas, attested to the valuable and timely insights he is able to reap by using Autotask's performance dashboard in conjunction with QuickBooks. For instance, the dashboard shows how many outstanding tickets are in the work queue; from there, Meek can make appropriate staffing decisions to ensure all projects get completed.
In addition, implementing a predictive analytics regimen boosts accountability. "Clients have conveyed that [with] having a fact-based or data-driven analysis system as opposed to an anecdotal system, people are more accountable for their results," Maisel said.
But that hinges on how integrated predictive business analytics becomes in an organization's decision-making process. Both Maisel and Rob Livingstone, principal and owner of Rob Livingstone Advisory PTY LTD., based in Sydney, Australia, emphasized this as a critical consideration for companies just starting out with a predictive analytics initiative.
"That's probably the first and foremost point to really work out: Does the organization have the ability to make appropriate decisions with the output from predictive analytics?" Livingstone said. "You can spend a lot of money doing predictive analytics, you can produce beautiful models, but if the decision-making capability isn't there," the results lose their impact.
Two impediments to predictive business analytics that also revolve around people are a lack of the appropriate analytical skills and a time crunch.
To the latter point, Maisel said some organizations already struggling with overloaded workers perceive the implementation of predictive analytics as unnecessary extra work. But in his opinion, this is the wrong outlook.
"Too much of the time of a finance organization is spent in capturing and correcting data as opposed to analyzing data," Maisel said. Predictive business analytics then is "not an add-on; it's a better prioritization of your resources and creates much more value as a business partner with your operating groups."
Skills are also a relatively scarce resource. "Accountants and finance folk for the most part don't necessarily have skills in data science," Livingstone said. "Data science is very, very different from doing modeling in Excel."
Developing new skills is important, but not everyone is going to become a data scientist, as Christopher Iervolino, research director at Stamford, Conn.-based Gartner Inc., pointed out. And this brings to bear a technological impediment.
"These capabilities need to become easier to use and more adoptable. We're not going to be throwing mathematicians at every problem," Iervolino said. "In addition, easier to use can mean easier to use successfully or unsuccessfully. Business knowledge is required in conjunction with appropriate training to help build a data-driven decision-making culture."
And getting the right data in order in the first place can be a herculean feat.
"Trying to cast a net to aggregate the various data sources is no trivial exercise," Livingstone said. "So when you're doing predictive analytics, one of the key things is [finding] all data sources that are appropriate, accurate and relevant."
Fragmented or not automated systems add to this difficulty, Maisel said.
Getting started with predictive business analytics: First steps
While it's clear that there are several real hurdles standing in the way of predictive business analytics, that doesn't mean launching a predictive business analytics program is impossible. These experts also weighed in on good places to start.
Iervolino suggested that organizations first understand their strategic priorities and then target the necessary data and technology.
"I recommend starting with strategic priorities and then exploring the data and technology available to support them. This will help keep the effort relevant," he said. "Untapped data sources may exist. The increased automation of internal processes and availability of additional micro and macroeconomic data sets can provide new sources of information that can be leveraged; for example, as leading indicators within existing KPI [key performance indicator] metric frameworks."
In Livingstone's view, organizations should start by assessing the decision-making structure.
"The very first thing before they lift a finger and invest in analytics is to work out what decision-making processes they have," Livingstone said. "What is the capability of the executive team in that organization to synthesize [data] and actually act?"
More on business analytics
Discover how business analytics aids decision making
Learn about the pros and cons of business analytics tools
Get tips to prepare for an advanced analytics project
In addition to determining how predictive analytics will integrate into the management review process, Maisel also suggested a few other steps.
"Figure out where you're going to make the impact," he said. "Why is this compelling, [and] what's the scope -- how broad do we want to make it?"
Maisel also pointed to the Professional Accountants in Business Committee paper "Predictive Business Analytics: Forward-Looking Measures to Improve Business Performance," which lists seven guiding principles that should underpin predictive business analytics initiatives. These include ensuring data integrity, providing relevant, reliable and timely data for decision makers, employing accessible and user-friendly tools, and incorporating a mix of financial, nonfinancial, internal and external measures.
And companies don't have to go all-in right away if they're not prepared. The paper suggests using pilot or scaled phase-in implementation approaches depending on the company's specific circumstances.
"Start with a reasonable scope, demonstrate its value early on [with] short-term wins and clearly leverage your staff," Maisel said. "In most organizations, many people would prefer to be analytical and less transactional."