Everyone Knows that Ed Tech is Expensive. But is it?

Education technology offers the promise of better outcomes at lower cost. Lower cost? How can that be–networks and devices and software come at real expense.

If ed tech can enable these better outcomes while allowing a higher student:teacher ratio, and if the savings in personnel more than offsets the cost of the ed tech itself, then its a win on both fronts.

So what does ed tech cost?

For the past 20 years, I’ve served a town in the Boston Metrowest region as a Finance Committee member, a School Committee member, and a technology task force member. As such, I’ve had a front row seat to the implementation of ed tech in a small district (of fewer than 3,000 students).

Over the span of the last 10 years, the district has moved from a relatively ad hoc deployment of ed tech to a much more thoughtfully developed, streamlined, and integrated system.

  • The first step in this modernization was the development of a network infrastructure. A data center at the high school connects to the Internet and via optical fiber to the other four buildings in the district.
  • The second step was the distribution of laptop computers to the educators and the provision of professional development to support their use.
  • The third step was the rollout of a student one-to-one computing initiative. First at the high school level (laptops) and this year at the middle school level (Chromebooks), each student is issued a device.
  • Step four is currently in its early stages: the deployment of instructional technology on top of the infrastructure layer.

A reader might be forgiven for thinking, “Of course the leafy and relatively affluent suburbs of Metrowest Boston can afford an effort like this, but what about school districts with fewer resources?”

Before taking that stance, it’s worth examining what the effort outlined above actually costs.

The budget shows a technology expense just into 7 figures. Yes, that puts us in “million” territory. And this is only for a small district–what about large cities like Lowell or New Bedford or Worcester or Springfield or Boston?

Let’s dive into the numbers. The budgeted ed tech operating cost for the 2015-2016 school year comes in at $1.04M. Add an allocation of ed tech capital at $130k and the total approaches $1.2M.

Expensive, right?

Not exactly. Here’s how that $1.2M breaks down on a per student basis.

  • Ed tech staff: $180
  • Software: $35
  • Equipment: $200
  • Other expenses: $30
  • TOTAL: $445

There are several caveats worth noting. Were the district to expand one-to-one to the elementary level, costs will go up. As the district expands instructional technology, software will go up. So let’s allow for $600/student/year.

What does the district spend overall in a year? A bit more than $16,000 per student. As such, the future ed tech spend represents a bit less than 4% of the total budget.

Meaningful? Yes. But bank-breaking? No, especially if an outcome of the technology deployment allows a higher student:teacher ratio.


A promise partially fulfilled

learnlaunch2015The occasion of LearnLaunch’s just-completed 3rd annual conference, “Across Boundaries: Delivering on Edtech’s Promise” (Jan 23-24, 2015, at Harvard Business School in Boston) reminded me of a white paper I’d written while at Lexia Learning almost a decade ago.

I revisit that white paper here.

Breaking through the school productivity ceiling:
the promise of education technology

Unlike almost every other field of human endeavor, education has not seen productivity gains, and as a result finds itself in a perpetual funding crisis that fundamentally limits its ability to improve student learning.

“Productivity” in the context of education merits a bit of elaboration. I originally wrote it in the literal numeric sense of students per teacher. To be sure, a graduating high school senior today knows considerably more than he or she would have a century ago.

Per pupil expenditure is an important measure of school productivity that—while admittedly not measuring relative education quality—has not increased over the last century or so. Take the teacher of the early 1900s, adjust that teacher’s salary for inflation, factor in class sizes that if anything have gone down, and acknowledge that productivity has at best been flat.

I don’t have a source at my fingertips, but I recall coming across statistics showing average high school class sizes in the 1900s being on the order of 40+ versus sub-25 today. And this article from the right-leaning Newsbusters uses Census Bureau data to show a 40% decrease in class size (and therefore productivity) since 1960.

Education technology that will deliver a combination of instruction, practice, and assessment has for years offered the promise of breaking the productivity ceiling.

By “combination,” I was referring to what I call “closed loop” adaptive instructional technology in which the technology itself serves up the next piece of learning based on how the prior piece of learning was received. This is in no way an argument for the removal of the teacher from the loop.

One model is to allow students to work with engaging and effective technology in a relatively unsupervised setting while freeing up a smaller number of teachers (through normal attrition) to work with smaller groups where their expertise is most valuable.

 As an aside, I see educational technology as being the combination of “platform” technology (network infrastructure, hardware, operating and learning management systems, and office and other productivity tools) and “instructional” technology (software that instructs).

Advances in the pedagogy underlying software content coupled with the evolution of hardware and infrastructure allow that promise to be realized within the next five years for any school district willing to pilot and then implement a technology-centric educational system.

Five years would have been 2010, and sadly, we didn’t quite get there. That said, in 2010, numerous examples of instructional technology existed, and the number has only increased since. Upon re-reading, “technology-centric educational system” was poorly phrased. “Technology-powered” or “technology-enabled” would have been far better.

These school districts have the potential to substantially alleviate their funding crises for at least the near and perhaps into the medium term.

In brief, here’s the model. If a teacher with a class of 25 students using conventional approaches can use technology to get better educational outcomes with a class of 30 students, there would be a cost reduction of 20% (less the added technology cost, of course). To put the numbers in perspective, consider that districts in the Metrowest area outside of Boston spend more than $16,000 per student, and that on the order of 2/3 of this cost is in the form of teacher compensation. Call that $10,000. A 20% savings would be more than enough to cover the cost of platform and instructional technology.

To realize this success, a relatively narrow education technology path is necessary.
a. The technology path cannot add operational costs in the form of IT staffing.

Actually, that’s not really the case–the cost of added IT staffing just needs to be factored into the overall cost picture.

b. The evolution of networks and network security are making school software implementations more rather than less complex, driving up the need for unacceptable IT staffing.

Given the many platform and instructional technology options facing schools, complexity is certainly part of the landscape. As the next point outlines, however, the web-based nature of these options dramatically reduces the potential IT burden.

c. Web applications (including “light” downloads with Web-stored data) get away from the expensive and not always reliable network model by putting the implementation burden on the software publisher and not the school district or its IT staff.

d. To date, the relative unavailability of reliable-enough Internet connectivity has slowed the movement from network-based applications to web-based applications.

Contradicting the next point, connectivity continues to be a problem.

e. However, the connectivity problem is diminishing reasonably quickly over time.

f. The viability of Web applications will enable an education technology-driven school model that will improve outcomes and decrease cost. Print and software publishers that fail to embrace this model will lose importance over time as measured in single digit years.

 I probably wasn’t too far off here. The big publishers, having seen their textbook “empires” put at significant risk, have been investing heavily in educational technology. Whether this investment (in the form of internal development and external acquisition) will be enough to allow them to maintain their market share remains to be seen.

Learn Launch conference delivers fantastic conversation on digital learning

Boston-based LearnLaunch‘s second annual conference (agenda here, but perhaps not permanently–also here as a PDF) brought together educators, students, entrepreneurs, industry participants, researchers, and other thought leaders to brainstorm on the question “How can Edtech scale student success?

The event was first-rate. LearnLaunch adroitly used the late Friday afternoon and Saturday calendar slots to allow attendance by many luminaries from the Boston education scene and numerous locales beyond. I came away with new connections, new ways of thinking about the challenges facing education and the approaches that might overcome these challenges, … and new ideas for blog topics. Over the next few weeks, I hope to explore some of these ideas.

My short answer to the conference question: Edtech can scale student success by delivering effective offerings to schools equipped to receive those offerings. In my most recent post, I touched on the second half of this success–schools needing the infrastructure to implement effective offerings. I’d like to use the current post to address the offerings themselves.

What makes for a better educational offering? I’ll argue that it’s a product or service that improves outcomes, reduces cost, or both. With digital educational offerings, there’s a clear path to doing both.

IMPROVING OUTCOMES: We learn best when we’re engaged. A key element of engagement is being appropriately challenged–not bored, not overwhelmed. In a classroom of 25 students, it’s impractical for a teacher to have every student engaged. Impractical, that is, without the right tools. Well-designed software offers the promise of providing this challenge-level engagement, allowing the teacher to take a step back and apply his or her talents where they’re most needed.

Are critical developments needed to create software delivering on this promise across the spectrum of student ages and curricular needs? No doubt. But are there examples of where this software exists today? Also no doubt–to wit, see Lexia Lexia Reading Core 5 (reading), Symphony Math (math), Harvard’s EcoMUVE (science), and McGraw-Hill’s MuzzyLane-developed Government in Action (social studies), to name just a few.

REDUCING COST: Teachers spend their time in many ways: preparation, collaboration, instruction, assessment, professional development, communication, and tending to various supervisory and administrative tasks. As it has in so many other fields, technology should be able to free up teacher time in most if not all of these areas.

Consider three possible examples of technology-enabled time saving:

Example 1: An elementary school teacher’s 2nd grade class increases from 20 to 25 students. Rather than dilute his time, the teacher rotates groups of 5 students through sessions with reading or math skill development software.

The software presents each student individually with instruction in the form of video or animation. The student practices what she has learned. In the background, the software assesses her responses, diagnoses her needs, and presents her with the next piece of instruction … all at a pace and challenge level matched to her needs.

While each group of 5 students interacts with the software, the teacher continues his work with the remainder of the class. The number of students served increases by 25% with no added staff. In effect, time is saved.

Example 2: A middle school teacher prepares for professional development. Instead of the one size fits all brand of in-service PD (and sometimes the need to travel to receive it), however, she accesses the course content online at a time and place that fits her schedule. Like the student in Example 1 above, she learns at a pace that works for her. A savings of even half an hour a week adds up to valuable time over the course of a school year. But her PD isn’t just available outside of class. It’s also available in short “sound bites” she can hear or read in the classroom in the context of a specific student’s need.

Example 3: A high school English and Language Arts teacher starts his year with a new energy. Why? A new system helps him grade the 100 papers submitted by his 4 classes of 25 students for every assignment. Students check their papers into the school’s learning management system (LMS). The teacher checks his dashboard, sees that 5 students haven’t turned in their work, and triggers a reminder message, avoiding having to track down missing items.

The teacher runs a first-pass grading utility that assesses the papers for spelling and grammar. He takes a second pass himself to judge each student’s style and quality of writing, making comments inline as he goes. Another push of a button and the papers are returned to the students. Even five minutes saved for each of the hundred papers frees up hours of teacher time.

How can Edtech scale student success? The offerings and implementations above show just several of many paths to that success.

Imagine the perfect digital learning offering

one-to-one computingFew are bigger proponents than I am of the application of digital technology to change the K-12 educational landscape. There’s a strong case to be made that technology will enable us to get better outcomes while reducing cost.

Let’s imagine that an education company has developed a suite of digital learning products spanning all grades and subject areas that does just that–learner understanding increases while the cost per learner drops.

Houghton Mifflin Harcourt’s CEO, Linda Zecher, is certainly not alone among education company executives pursuing this vision. In an August 2013 Boston Business Journal article, she’s quoted as setting the goal of having 50% of the company’s revenue come from digital products in 2015 (up from 30% in 2013).

Setting aside the question of how digital revenue will be measured (if a product has a digital component, will its entire revenue be counted as digital?), I applaud that goal.

What I haven’t heard mentioned, however, is what I think is a major roadblock to achieving the goal, especially by 2015. Put simply, a digital education requires students to have access to a device at any time. In short, that means one-to-one computing.

The natural question to ask, then, is what fraction of schools have implemented one-to-one computing? While this data is getting a bit old, MDR/EdNet Insight’s 2011 “State of the K-12 Market” (Part III) finds that based on a survey of 300 districts, only 11 percent report having at least substantially implemented one-to-one (pp. 19-22).

  • Full implementation (5 on a scale of 1 to 5): 3.8% of districts
  • Substantial implementation (4): 6.9%
  • Partial implementation (3): 15.2%
  • Partial implementation (2): 23.8%
  • No implementation (1): 50.3%

In sum, education companies such as HMH may be ready (or readying) for digital, but their customers aren’t yet keeping up–half have no implementation whatsoever. The education companies that succeed will be the ones that offer hybrid print/technology products and services that serve districts both with and without one-to-one implementations as they collectively navigate the transition to digital.