I have gotten a lot of very nice compliments in the last 24 hours about the first post in this two-part series. I do want to emphasize that a huge portion of the value in that post comes from the great survey work that Casey Green does. All I did was tease out a few implications and make a few new graphs (which, I admit, were very pretty). If you want to see analysis like this continue in future years, then support the Campus Computing Project.
In this next part, we're going to go "off-road" a little and see what we can figure out in areas where we don't have quantitative data that is as solid as Casey's. As I wrote in the previous post in this series, there are roughly 875 WebCT and ANGEL customers who will have to migrate to a new LMS in the next few years, in an environment of strong budget pressures. This creates an atmosphere in which more and different schools may be in play than has typically been the case. But evaluating options and choosing to move are two different things. What happens in the next few years is likely to shape the LMS landscape for years to come, at least in North America.
Let’s see if we can read the tea leaves.
More on Market Share Numbers
One of the (many) reasons that figuring out what’s happening is so hard is that we don’t really have good, granular data on LMS market share by LMS platform, and the data that we do have appears to be contradictory on the surface. The numbers we get from Campus Computing are very good as far as they go, but they break down market share by company rather than by platform. We have no way of knowing from the survey data, for example, how many WebCT Vista customers there are versus Blackboard Learn Basic, etc.. Blackboard used to break out Blackboard Learn Enterprise licenses versus all other LMS licenses in their public reporting, which was better than nothing, but they stopped doing even that much last year.
On top of that, the aggregate market share numbers we get from Campus Computing seem to be in tension with the customer retention numbers that Blackboard reports. If you strip out the noise from acquisitions, Campus Computing’s survey numbers suggest that Blackboard is losing 6% to 8% organic market share annually. But Blackboard regularly reports customer retention numbers in the 90% to 92% range. Can both of these be right?
Actually, they can. When a company reports customer retention numbers, that’s usually an indication of churn, i.e., how many customers are moving around among the competitors. It matters because there is a cost of sale; every new customer costs money to win over. Therefore, the more customers you hold onto, the fewer new ones you have to recruit to stay even, the more profitable you can be. But Blackboard’s situation may be somewhat different. As far as I can tell, their retention rate reflects not churn, primarily, but almost pure attrition. I am not aware of Blackboard acquiring any new higher education customers for their LMS in North America in recent years, and none of the experts that I asked in preparing to write this blog post were able to cite any either. This doesn’t mean that there haven’t been any, but it suggests that there probably haven’t been many. The company may be picking up a few new customers overseas and a handful of domestic K12 and government/corporate customers, but for the most part, it looks like that 8% to 10% of the LMS customer base lost every year is not being replaced. An 8% to 10% attrition rate at their market share level is roughly compatible with a 6% to 8% loss of overall market share. Furthermore, it’s not entirely clear how Blackboard has been defining retention. If a customer drops a Blackboard-owned LMS but continues to license the Blackboard-owned Xythos content management product, does that customer count as retained? From a cost-of-customer-acquisition perspective, it would be perfectly reasonable for a company to count that customer as retained. But if you’re trying to understand what’s going on in Blackboard’s LMS business, then that sort of accounting wouldn’t tell you a whole lot. More recently, Blackboard management has referred specifically to “revenue retention,” although it’s not entirely clear whether that has always been their definition, nor is it entirely clear how the company derives such a metric. Again, since neither Blackboard nor Campus Computing breaks out customers by the various platforms Blackboard owns (and, in Blackboard’s case, they no longer even consistently break out LMS licenses from other licenses), we don’t have much in the way of hard numbers to help us reconcile these various bits of data and management statements that we’re getting.
At any rate, we have about as good a picture of the recent past as we’re likely to get. There are three questions going forward. First, are the kinds (and sizes) of customers that migrate off of Blackboard going forward different than the kinds of customers who have done so in the past? Second, will the trend away from Blackboard-owned platforms accelerate or decelerate? And finally, will the proportion of departing Blackboard customers going to each platform alternative change going forward?
Kinds of Customers
As I wrote at the top of the post, we’re getting into an area in which we have very little hard data. But from what Blackboard has said on their earnings calls, it appears that the majority of customers they have lost in the past have been WebCT CE and Blackboard Learn Basic customers. By and large, these were low contract value customers. My guess is that they also have tended to be small in terms of student population, but I don’t have the data to prove it. What I do know is that in November of 2007, then-Vice President of Investor Relations Michael Stanton told me that the company may have lost “one or two” Vista customers since WebCT had been acquired in February 2006. Vista was WebCT’s enterprise platform, where it’s larger and higher contract value customers tended to be. And at that point, they were staying.
There are signs that this the state of affairs may be changing, though. Consider:
- University of North Carolina, Charlotte (24,700 students) has decided to move from Vista to Moodle in 2011.
- North Carolina State University (31,000 students) has also decided to move from Vista to Moodle in 2011.
- Wilfred Laurier University (16,000 students) moved from Vista to Desire2Learn last quarter.
- University of Delaware (19,391 students) just completed a move from Vista to Sakai last quarter.
- University of Minnesota (51,721 students) is moving from Vista to Moodle in 2012.
- The Utah Education Network (109,000 college students plus 40,000 K12 students and teachers) is moving from Vista to Instructure Canvas in 2012.
These are just anecdotal, of course, but they raise the possibility of a trend. Particularly important to note is that Vista supports multi-tenant hosting capabilities beyond what Blackboard Learn currently offers. (Learn is reportedly going to come into parity in this regard with version 9.2.) This is particularly important for multi-campus systems or consortia that host their LMS internally. That means these large customers are not likely to have made the move to Blackboard yet. So, for example, UMassOnline (77,108 students), which provides LMS support for distance learning to five University of Massachusetts campuses plus a handful of other colleges throughout the state of Massachusetts, is on Vista (and is in the middle of an LMS platform evaluation process). These multi-campus systems and consortia are the ones to watch very closely.
The ANGEL demographic is somewhat different. Its strengths were in community colleges (13% market share in 2009, according to Campus Computing Project) and private four-year colleges (10.3% market share in 2009). Again, these categories don’t necessarily indicate contract size. For example, the SUNY Learning Network hosts ANGEL centrally for something like two dozen community colleges. And Monroe Community College—a member of the SUNY Learning Network—has about 19,000 students by itself. Miami-Dade College, another ANGEL customer, has about 170,000 students. The other outliers that I know of in terms of large customers (regardless of segmentation) are Penn State, with 87,000 students, and Michigan State University, with 47,278 students. (SUNY and Penn State are both in the process of respective LMS reviews; there’s no public data I could find on whether Michigan State or Miami-Dade begun any such evaluation at this time.) But overall, my sense based on anecdotal evidence is that the ANGEL customer base probably doesn’t look very different from the CE customer base, in both campus size and contract value. With a handful of exceptions, ANGEL customer attrition could look a lot like the Blackboard Basic and WebCT CE customer attrition that Blackboard experienced from 2006 to 2010—assuming that they leave in the same proportions.
Numbers of Customers
In addition to the kinds of schools that could potentially migrate, the number of schools that migrate from 2011 to 2014 may be different than from 2006 to 2010. We're not going to find that out based on the number of forced migration candidates, though. The number of combined WebCT and ANGEL customers that need to migrate in the next four years is roughly equal to the number of WebCT customers that migrated off of the legacy platforms in the last four years so, if everything else holds equal, the rate of attrition shouldn’t change that much. What we don’t know is the degree to which economic pressures combined with the increased perception of viable alternatives in the market—or, for that matter, changes that Blackboard makes such as improving customer service, which has been a big push for them of late under Ray Henderson’s leadership—will change the rate of attrition.
Once again we don’t have hard data and are forced to look for anecdotal evidence for the answer to this question. But one kind of anecdotal evidence that would be particularly potent would be customers that are not on legacy platforms moving away from Blackboard. As far as I can tell, the historic attrition of Blackboard Learn Enterprise customers has been very low. Again, customers don’t tend to move from their LMS unless they feel they are forced to do so. If mainline Blackboard Enterprise customers are moving, it likely means that either the economics are driving them or they have had such catastrophic product or customer service failures that they don’t think they can afford to stay with the vendor.
Are there any data points indicating that that this kind of attrition is happening? Well, there are a couple that I know of:
- University of North Carolina Chapel Hill (28.916 students) are moving from Blackboard to Sakai in 2014.
- Miami University of Ohio (20,126 students) is moving from Blackboard to Sakai in 2011.
- CSU Long Beach is in the process of moving from Blackboard to Desire2Learn over the current academic year.
- The tri-college consortium of Bryn Mawr, Haverford College, and Swarthmore (combined student total of about 4,600 students) is moving from Blackboard to Moodle in 2012.
- The 14 schools in the Penssylvania State System of Higher Education, or PASSHE (119,513 students) was on a mix of Blackboard, WebCT, and eCollege—but mostly Blackboard—and they recently switched to Desire2Learn. (I’m not certain how many of those Blackboard campuses were on Enterprise, so take this one with a grain of salt.)
This is obviously not enough data to establish a trend. But any defection of a Blackboard Enterprise Learn customer is significant because there is no forced migration and because the barriers to exit are high. To gauge whether there is an acceleration of customer migration starting, I would watch for RFPs coming from Blackboard Learn Enterprise customers, and I would watch schools like Brigham Young (26,928 students), which has been heavily involved with co-designing and piloting Agilix BrainHoney, and Arizona State University’s (70,440 students) deal with Pearson LearningStudio (a.k.a. eCollege). If these schools begin to move to other platforms as their main LMSs, it would be a strong sign that WebCT and ANGEL customers are also likely to move to other platforms in numbers that are higher than in the past.
It’s harder to identify a leading indicator that might suggest the attrition rate is going to be lower. You’d have to monitor school announcements as they come out. If more than 50% of the announcements from WebCT and ANGEL schools are for upgrades to Blackboard Learn Enterprise 9.x, and particularly if the big Vista-using consortia start announcing moves to 9.x in numbers, those would be signs that Blackboard may be lowering attrition rates from their historical norm and holding onto their high-value customers in particular.
In terms of where these customers end up if they leave Blackboard, we have very little data to go on beyond what I covered in my last post. It’s possible that the Vista and ANGEL customers will have different platform preferences than the CE and Blackboard Basic customers have had, so the numbers we have been seeing in terms of relative market share gains among the contenders could change. And, of course, there’s always the possibility of the breakout success of a player other than the big three taking a few percentage points of the market share pie. The 2011 Campus Computing survey will be particularly important to watch, I think, in telling us the degree to which the trends are holding.
The Bottom Line
Here’s what we know about what’s happening in the LMS market, based on the analysis of this post and the previous one:
- The data confirm that the LMS market is very much in transition.
- Projection of current trends over the next four years suggest a world in 2014 in which at least four platforms have comfortably over 10% market share and no platform dominates all others.
- There is no evidence to suggest that open source is either being confined to small overall market share or on its way to dominance of the market in the next four years.
- There is some evidence that schools within particular market segments may be predisposed to choose open source, while schools in other market segments may be predisposed to avoid it.
- There is good evidence that preferences for particular LMS platforms trend strongly by market segment.
- Within those broad outlines, there are a number of uncertainties related to the economy, the largely unknown switching preferences and tendencies of Vista and ANGEL customers, and the actions of the individual LMS vendors and open source projects that could change the trends somewhat from what we’ve seen in the past.
- We don’t have any evidence yet to suggest that any such changes are likely to have a very big impact on the overall projections for market share in 2014, although they could have significant impact on individual LMS players where winning or losing a few large deals could have an impact on their businesses, and possibly on the market share numbers within individual market segments.
- Anecdotal evidence suggests that there may be room in the market for one or more LMSs outside of the big four to pick up a few points of market share over the next four years.
- Given the sunset dates for WebCT and ANGEL products, most schools that are likely to participate in this change wave will start their evaluations within the next 6-18 months. That means the LMS versions that are available in calendar year 2011 are the ones that they will be evaluating, by and large.
Of course, four years is a very long time in the technology world. I expect the context of the market to be very different by 2014. Between movements toward more personal/informal learning environments, the big changes that are happening in the textbook industry as content providers and delivery platform providers collide, and the breathtaking pace of innovation that continues in the consumer web market, I strongly suspect that we’ll see a wave of new software that will begin to displace the classic LMS. My guess is that it will be mostly co-existing with current-generation products over the next few years, but by 2014 we may see it beginning to change the whole picture for educational technology infrastructure in some fundamental ways.
Buckle up, folks. It’s going to be an interesting ride.