This year students at Colorado State University will have their progress toward degrees tracked by technology from a company called CollegeSource. The likelihood of their encountering academic difficulty will be flagged for academic advisers and resident advisers by an online product from a company called Campus Labs. And they will receive text alerts about unsatisfactory grades via a mobile app from yet another ed-tech company, Ellucian.
Students in three majors will also be assessed on their chances of succeeding in a course, on the basis of an analysis of data from thousands of previous Colorado State students who had earned the same grades, thanks to technology from the Education Advisory Board. Meanwhile, intercollegiate athletes who skip too many classes will be notified through Facebook by a company called GradesFirst that they’ve been scheduled for tutoring. And all these arrangements don’t even reflect an experiment in evaluating student progress in courses, using an analytics tool sold by Blackboard. The university ditched the experiment in the spring after realizing that professors weren’t using the learning-management system uniformly for that purpose.
Colorado State, with 27,000 students, has become a test bed for an array of companies selling “personalized education” products. But its reliance on so many vendors has happened more by necessity than intention. “We realized no one system is the answer,” says Gaye DiGregorio, executive director of the university’s Center for Advising and Student Achievement. “When we started this, that’s what we were looking for.”
Though campus officials are confident about the potential for improving student learning through the use of these products, which cost more than $900,000 a year, they are also frustrated. “It’s not as streamlined or as smooth as we would like it,” says Ms. DiGregorio.
Campuswide student-information systems became commonplace in the 1990s, while learning-management systems dominated the IT scene in the early 2000s. Today in ed tech, the era of personalized education is taking hold. Along with established companies, like Blackboard and Ellucian, that offer personalized-learning products, the field includes dozens of companies backed by hundreds of millions in investments.
The meaning of “personalized education,” however, is still rather fuzzy. The personalized-learning market encompasses companies as diverse and hard to classify as Starfish Retention Solutions, which flags students who miss assignments and allows professors to send them “kudos” when they do well; Acrobatiq, which analyzes which instructional approaches work best for which students; and Civitas Learning, whose faculty-friendly “heat maps” give professors a quick visual representation of how their students are doing.
The offerings from those companies and others, including those whose products Colorado State is using, hold enormous promise for helping more students succeed in college. But the frustrations at Colorado State and other colleges that increasingly rely on ed-tech offerings aren’t likely to dissipate soon.
The companies are taking shape in what observers like Adam J. Newman of Education Growth Advisors, a consultancy, calls “an immature market.” For colleges, that can spell confusion. Analysts note that many institutions are still unaware of what’s out there and unsure of how to integrate products from different companies.
On the part of companies and colleges alike, says Mr. Newman, “there is a woeful underinvestment” in ensuring that the products are designed and deployed in ways that are appealing to those whose buy-in matters most: the professors. “Solutions have to win the hearts and minds of faculty,” he says.
Colleges’ newfound interest in personalized-learning products and services also highlights the potential for unrealistic expectations. “People are jumping to quick conclusions” about what the products can do, says Mark David Milliron, a former official with the Bill & Melinda Gates Foundation and a co-founder of Civitas, which makes predictive-analytics tools. “But there aren’t three magic bullets that fix everything. This is art and science.”
Michael Feldstein, a partner at MindWires Consulting and co-publisher of a blog called e-Literate, says “personalized learning” is a misnomer and a dangerous one at that. It’s an emotionally charged term that may promise more than it can deliver because none of these products are really about helping students develop as people, he says. And in a highly stratified higher-education system, he worries that the term overpromises. “That gives the illusion that this is somehow equalizing the education,” he says.
To be sure, it’s hard to know what people mean when they talk about personalized learning. The term is often applied to companies like GradesFirst and Starfish, which focus on retention and advising, as well as to companies like Acrobatiq and Smart Sparrow, which are oriented more toward instructional tools and courseware but also offer tools that show how students are faring.
Right now it’s a challenge to even compare products, because they’re not aligned, says Terri-Lynn B. Thayer, a research director in higher education at Gartner, an IT-research company. “It’s like apples, oranges, and pears,” she says. But the companies share one unifying trait, she notes: “Data analytics is the underlying common denominator.” She expects more innovation by companies, more attention from colleges, and more clarity in how companies define themselves over the next two years. “I think it’s going to move quickly,” says Ms. Thayer, who was formerly chief information officer at Brown University, “because of the urgency in our industry.”
Players in the personalized-learning business approach their subject from varying perspectives because of their varying origins. Starfish, for example, was founded in 2007 by executives with experience at other companies in the higher-education market, like Blackboard and the Ellucian precursor Datatel. It emphasizes connectivity, working to create products that tie in to diverse data and course platforms and that appeal to faculty members who use them. Starfish has a campaign aimed specifically at encouraging faculty participation. It includes emails from the college president to the faculty, says David Yaskin, founder and chief executive.
Starfish, in use on more than 240 campuses, manages data on the activities of more than three million students. “Different kinds of students get different levels of support,” says Mr. Yaskin. For example, one college creates extra alerts for athletes, military veterans, and first-generation college students.
Big campuses would seem to be the target market for such products, and that is the case for Starfish. But the company’s clientele also includes institutions as small as Paul Smith’s College, in the upper Adirondacks, with about 1,000 students. Now beginning its fifth year with Starfish, the college says its retention rate has improved from a percentage in the low 60s to about 70. It calculates that during the same period, its retention efforts (which go beyond Starfish’s services) have saved it about $2.6-million in what would otherwise have been lost tuition. “It’s like a second endowment,” says Phillip A. Taylor III, a dean there.
Bob Seidenstein, who recently retired from the college after teaching writing for 41 years, says he was dubious at first, especially since the small-college culture already fostered a supportive environment. But the self-described “computer illiterate” came to appreciate the ease of the system, which allowed him to chide students for failing to complete class assignments without directly confronting them. “That’s the thing with teaching—you have to get them to get their own internal motivation.”
For other companies, faculty reaction isn’t as important. GradesFirst began in 2006 as a tool for athletics departments to keep athletes on track for their advising sessions, but it has since expanded, a reflection of the way concern about academic progress has gone more mainstream. Today, says Mario Moore, founder and chief executive, athletics accounts for only about a quarter of the company’s business.
GradesFirst doesn’t even hook its system into the course platform, pointing to survey findings that fewer than one-third of professors use such systems to take attendance. Nor does it rely on the predictive analytics that other companies use to identify behavior that could indicate a student is in academic trouble. Professors can use the system to flag at-risk students, but it is primarily designed to tie into the advising and support systems, by keeping track of students who miss a tutoring session, for example, and then ensuring that they get signed up for the next one. “We focus on what happens next,” says Mr. Moore.
Some personal-learning companies are more oriented toward instruction. Acrobatiq, a for-profit company that Carnegie Mellon University spun off from its grant-financed Open Learning Initiative (OLI) about a year ago, specializes in creating courses grounded in the science of learning. It also develops tools to track the effectiveness of those approaches based on students’ class performance. It then incorporates its findings into “dashboards” that can show instructors what’s working and what’s not, and provide students with personalized practice sessions based on the proficiency they demonstrate during the course. The information also suggests ways to redesign the presentation of material in future versions of the course.
Acrobatiq has redesigned five courses that were created by OLI and is developing 24 additional ones for Western Governors University and other clients. But its scope is evolving as it moves toward becoming a tool company that will make such personalized approaches available to other course developers, says Eric Frank, the chief executive. He came to Acrobatiq after founding Flat World Knowledge, a digital textbook and courseware company. (While OLI’s founding researcher, Candace Thille, has moved to Stanford University, OLI continues to pursue new research in learning science, with $500,000 of retained grant money, $500,000 in new grants, and other support from Carnegie Mellon.)
Civitas, too, sees itself as a tool company, one focused on creating apps that take advantage of the growing mountains of data that colleges collect on their students. “Every day the digital footprints are getting bigger,” says Mr. Milliron, and the technology is finally in place to take advantage of that. “You could not do what we’re doing five years ago.”
For example, the “heat map” created by Civitas to give professors a quick image of student performance is updated as frequently as every four to six hours, based on data coming in from sources like the student-information system, which might indicate if students have dropped other courses, and the learning-management system, which can show if students’ logging-in patterns have changed, which could be a sign that they are cramming rather than planning.

Civitas
A tool made by Civitas Learning can help professors keep track of how each student is performing.
Founded in 2011, Civitas spent the past year working with six “pioneer” institutions at no charge. Recently it announced it was expanding that base to 34. Its focus is on designing tools “that get the right data to the right people in the right way,” Mr. Milliron says.
Along the way, Civitas has come to realize that what works for some institutions may not work for others. At one of its pioneer institutions, the University of Maryland University College, administrators initially included “attendance triggers” in the academic-monitoring system but found that they weren’t reliable indicators of success for a student body that includes many military members and older students. The alerts “were bothering students and making them mad,” he says.
Ms. Thayer, the Gartner analyst who spent decades as an information-technology leader at Brown, says many colleges would be wise to remember that lesson as they embrace personalized learning. “As an industry, we’re still trying to understand what makes students successful,” she says. “We don’t really know what the algorithm is for student success.”