Predictive analytics is experiencing what David Menninger, a research director and vice president at Ventana Research Inc., calls “a renewed interest.” And he’s
In September, Hurwitz & Associates, a consulting and market research firm in Needham, Mass., released a report ranking 12 predictive analytics vendors that it views as “strong contenders” in the market. Fern Halper, a Hurwitz partner and the principal researcher for the report, thinks predictive analytics is moving into the user mainstream. She said its growing popularity is being driven by better tools, increased access to high-performance computing resources, reduced storage costs and an economic climate that has businesses hungry for better forecasting.
“Especially in today’s economy, they’re realizing they can’t just look in the rearview mirror and look at what has happened,” said Halper. “They need to look at what can happen and what will happen and become as smart as they can possibly be if they’re going to compete.”
While predictive analytics basks in the limelight, the nuances of developing an effective program are tricky and sometimes can be overwhelming for organizations. But the good news, according to a variety of analysts and consultants, is that finding the right strategy is possible -- even on a shoestring budget.
Here are some of their best-practices tips for succeeding on predictive analytics without breaking the bank:
Create a blueprint. Before you get started, do an assessment of your organization and create an initial blueprint for the deployment. That means evaluating internal tools, talent and data while also vetting the vision for what a predictive analytics program could help your business do.
Eric King, founder and president of The Modeling Agency, a Pittsburgh-based consulting firm that focuses on analytics and data mining, said many organizations could benefit from the assistance of an outside consultant when putting together a project assessment, which he estimates would typically cost between $20,000 and $30,000.
“I would suggest they do this under the guidance of a seasoned data mining consultant,” King said, “because it’s far less than A, purchasing a major commercial software package; B, failing at your first pass and having to reset and start over; or C, failing, deciding it didn’t work and realizing two years later that you have to do it anyway.”
While King believes a guide in the form of a consultant is an invaluable resource for businesses in the planning phase, he noted that his firm follows the Cross Industry Standard Process for Data Mining, a public document he describes as “a cheat sheet,” when it’s working with clients.
Build and test a prototype system. That means starting with a question, and analysts said to select one that you already have the data required to answer. In addition, they recommended, choose a question that likely will have an impact on your organization’s business operations.
“Maybe you don’t swing for the fences and try for a killer application that would make a huge difference in your company,” said John Elder, CEO of Elder Research Inc., a data mining and predictive analytics consultancy in Charlottesville, Va. “But you find something that would make a measurable, noticeable difference.”
Building the prototype may lead organizations to believe that they need to invest in software at this point, but King said not yet. “You don’t need to get the software that’s ultimately going to become part of the process,” he advised. “That can come later still.”
Instead, he and other analysts pointed to open source analytics tools such as Knime and Weka as possible choices, at least in the prototyping stage. Several also suggested looking into Software as a Service (SaaS) options.
“A number of vendors are selling predictive analytics services like market basket analysis or market segmentation via a SaaS model,” said Rita Sallam, a research director at Stamford, Conn.-based Gartner Inc. “You might be able to buy what you need, see how it works in your organization … [and] at least build the business case for the impact it could have.”
The prototype system also provides a real-world sample that can help shine a light on an organization’s culture, especially with corporate executives, and reveal whether there’s any internal reluctance to move forward on a predictive analytics program before you’ve invested large amounts of money in one.
“At the end of the day, we need to make the leap between the theoretical intention to use predictive analytics and the actual willingness to adopt -- and fund -- such techniques,” said Adrian Alleyne, director of market research for DecisionPath Consulting, a Gaithersburg, Md.-based firm that works with clients on business intelligence, analytics and data warehousing deployments.
Aim for sustainability. Look for ways to make a predictive analytics strategy self-sustaining. For example, when you’re creating the initial blueprint, John Lucker, who leads the advanced analytics and modeling practice at Deloitte Consulting LLP, suggests exploring potential predictive analytics applications and then categorizing them as short-, medium- or long-term goals.
“You want to calculate the return on investment for each of the different opportunities,” he said. “Then structure it in a way that the investment in the short-term projects yields an ROI that feeds the cost of the [larger] investments.”
Of course, all of an organization’s predictive analytics goals may not fit neatly into such an arrangement. If that’s the case, Lucker said a separate category -- what he calls the project parking lot of the predictive analytics road map -- could act as a holding pen for projects that aren’t easily self-funded until budget money becomes available.
Choose the technology path that’s right for you. When the time comes to select the predictive analytics technology that will be used as part of the deployment, businesses need to decide whether investing in software to run internally is the best choice or if other options, such as a SaaS model or hiring an analytics service provider, would be a better use of the available funds.
“If you’re a company that grows shrubbery,” Lucker said, “do you really want to be a great analytics company? Maybe you do -- maybe that’s important to you. But maybe you just want to be a shrubbery grower.”
In addition to software purchase and maintenance costs, finding people with the skills required to build predictive models and to structure, cleanse and normalize data for predictive analytics uses can be challenging -- and expensive. Turning to a service provider to manage the analytics process might be less costly than buying the tools and hiring skilled analytics professionals outright.
If an organization does decide that it’s important to build those skills internally, training workers who are already in-house to do predictive analytics could be a cost-effective avenue to consider, according to Elder. “The good news is people want to learn this and they can,” said Elder. “The bad news is you can’t learn it fast.”
Also, you don’t necessarily have to spend big money on predictive analytics software if you do want to buy technology and run it in-house. “While major vendors sell predictive analytics platforms that allow tight coupling back into operational processes, it’s also possible to get a low-cost tool -- $2,500 or less -- for one or two good analysts to use to impact a specific process,” Alleyne said.