Wednesday, November 07, 2007

Berger and Stevenson on K-12 Entrepreneurship

I'll second Dean Millot's recommendation of Wireless Generation's Larry Berger and David Stevenson's paper, K-12 Entrepreneurship: Slow Entry, Distant Exit.

Wireless Generation was founded in 2001, and seems to have a practical, useful, well regarded product aimed at a sector which has seen unprecedented growth in the intervening years, yet:

We have grown to 250 people, serving more than 2.5 million K-6 students, including most of the K-3 classrooms in New York, Chicago, Miami, Houston, Washington, D.C. and more than a thousand smaller districts. We are considered a preliminary entrepreneurial success story, but we are not quite profitable yet, and we remain vulnerable to political funding shifts, to the various ways that the culture of education is slow to change, and to competition from “Big Edu” (the three dominant educational publishing companies).

Waitasec... they "aren't profitable yet?" Whew. This market is even more brutal than I thought.

Berger and Stevenson explain that while there are a multitude of small education consultants (that is, smaller than them), the little guys don't generate much innovation:

The small business owners in education are not necessarily entrepreneurs. Many of them are building “lifestyle” businesses to keep themselves engaged after they retire from the school district. They are not attracting investment capital to drive accelerated growth and in many cases do not want to grow beyond a certain size. In their local community, they get to know the local landscape just like the local Big Edu sales rep, and they learn to thrive in that small pond. They are not entrepreneurs who want to innovate across the sector, to create value by redeploying resources, and to take high levels of risk for high levels of reward – enough potential upside to justify gambling years of low salary, high stress and uncertainty. The would-be entrepreneur scans the educational landscape, notes the crowded field of small companies and the rarity with which any of them cross the chasm to become big companies and starts looking at other sectors.

The bulk of the paper is an elaboration on their "top ten barriers to entry."

  1. The Education Sector Does Not Invest in Innovation
  2. Oligopoly
  3. Decentralization
  4. Vicious Sales Cycles
  5. Pilot Error
  6. No Return
  7. Viewing Teacher Time as a Sunk Cost
  8. Short-Lived Superintendents
  9. The Vendor Wall
  10. Start-Up Capital

I'd encourage the reader, in exploring the whole paper, to not slip into thinking of this list as ten manifestations of one problem (most likely, educators being generically resistant to change), but an interlocking web of problems. Short lived superintendents means true innovation is swamped in a constant churn of change. Education professionals cannot chose to invest in innovation on their own; their budgets are determined by political processes. There is not an easy solution like simply getting educators to think like businesspeople.

Berger and Stevenson do offer three possible paths forward, and I'm happy that "Create A Welcoming Climate for Promising Disruptions, Including Open Source Business Models" is one of them. They don't elaborate much on this idea, but I think I can.

One possible route is for open source processes to drive more innovation, and more distribution of innovation, by the "lifestyle businesses" described in the paper. In this kind of case I think of Plone Foundation and ZEA Partners. The Plone Foundation is made up of over 100 individual developers who contribute to the Plone content management system. Zea Partners is a non-profit business network for 25 companies building Zope, Plone and Silva systems. These developers and companies are probably more entrepreneurial than the "lifestyle businesses," but they aren't much bigger, and as far as I know, aren't aiming for rapid growth and an IPO. They're pretty much smallish web development shops. Working together through open source processes, they collectively innovate at a rate comparable to, if not faster than, a single big company

There is no reason that small education consultants can't create similar processes and organizations. There is also no reason that releasing open source software and explicitly cultivating such organizations should not be a goal of foundations, universities, etc. seeking routes to scale up innovations.

Also, Berger and Stevenson describe Big Edu's dominance of traditional distribution channels. It is important to try to start communicating the role of the free software community and free software distributions as an alternative distribution channel. While the web is in many ways the ultimate distribution channel for software, there is still something to be said for the integrated packaging systems of Linux distributions like Ubuntu (or, in this case, Edubuntu). At this point there still seems to be a gap between the innovative educational software being developed at universities and what ends up in K-12 oriented distributions like Edubuntu.

Case in point: the Ubuntu Developer's Summit was at MIT last week, including the prime movers of Edubuntu. Scratch, the second most interesting and important educational free software project of the year, also lives at MIT and is conspicuously not included in Ubuntu or any other free software distributions that I'm aware of at this point. I wonder if anyone from Edubuntu called anyone from Scratch to have a meeting, or vice versa (part of the problem here is that it isn't clear whose responsibility it is...). I don't know, but I'd be surprised if they did. The academics don't see the need for a new distribution channel, and the free software guys aren't used to initiating these conversations, especially when they've got more than enough work on their plate already, thank you very much. Nonetheless, these two camps need to be brought together; they need each other.

One final note. The paper says:

Draft: Please do not cite without permission from the author.

Which I interpret to mean, "don't cite this in an academic work" not "don't quote us on the web," which would seem kind of absurd since the paper has been posted on the web in its entirety. But consider yourself warned. Regardless, the whole thing is highly recommended, even if it was written for the execrable AEI.


Bill Fitzgerald said...

I felt dirty after I read that paper -- and not in a good way.

Lines like "The Reading First component of No Child Left
Behind created an unusual amount of liquidity centralized at the state level (about $200 million
per year) that did not already have a bureaucracy trained to spend it and that was distinct from
another $800 million per year for the districts. So we were able to visit state capitols and win
eighteen state Reading First assessment contracts, rather than having to visit the thousands of
districts those contracts comprised." (pg. 10) suggest that they were merely capitalizing on an opportunity in the market.

Follow that up with "Such legislation is almost unprecedented and certainly has not repeated itself for any of our other products." (pg. 10) and I am left filled with a sense of disgust. In this context, NCLB was just a business opportunity, as opposed to unsound educational policy.

Maybe they are too close to the market to see the connection between an underfunded mandate that rewards effective recordkeeping alongside a highly managed curriculum and the business opportunities for companies selling that boxed curriculum. Why are they surprised that the educational publishers are making a killing in this market? It seems that the crux of their complaint/observations are that they didn't have the sales force/political muscle to play alongside the big boys in the market.

Also, the fact that this was created for a presentation for the AEI is hilarious -- And you left out the best part, the title of the Event: "The Supply Side of School Reform and the Future of Educational Entrepreneurship"

I also enjoyed the fact that you couldn't actually link to any docs, but that you had to find the appropriate paper from among the docs listed on the margins. Great design.



Tom Hoffman said...

Wireless Generation is clearly in it for the money. They're a business, and how the K-12 education market works is what the article is about. What they personally think about NCLB as policy is outside of the scope of the paper.

What they provide is products and services that, from what I've seen, make it easier for teachers to collect assessment data and subsequently analyze and use it. Whether or not that is a good thing mostly depends on what data you're collecting and how you use it, but that's not so much their problem.

Doug Holton said...

Yeah Scratch has an MIT license, the PHET simulations from Colorado and Processing are GPL, and Alice has a permissive license. It would be nice if they were all in Edubuntu. I guess there might not be Ubuntu package versions though. That's probably the issue.

Tom Hoffman said...


Helping make packages is one of the few things the Edubuntu staff (which I think is still like, two people) and community can do. For that matter, Eric Harrison is frighteningly efficient at producing RPM packages for K12LTSP. Getting packages made is more a matter of setting up lines of communication more than anything else.

Simón A. Ruiz said...

Scratch is released by MIT, but I've not seen anywhere that it is released under an MIT license.

In fact, I've not seen any Linux-native release, source OR binary.

There certainly are a significant number of good, educational Gratis Software packages that are unavailable to us under Linux except through the use of WINE.

We have a 1:1 Math classroom where most of the applications that the teacher uses are emulated under WINE. And none of them are commercial, it's just they're only released for Windows.

Tom Hoffman said...


Scratch is essentially BSD-licensed. As it is based on Squeak, the source can be introspected from within the environment, although you currently have to sort of hack your way into it. They could (and should) make it easy. I'm not sure how meaningful it would be to release the code in a more standard form.

Similarly, as the Squeak environment is ultra-portable, David Thornburg hacked together a Linux port without much trouble. It would be fairly easy to make proper packages if they want to (or even if they don't, for that matter).