Friday, March 15, 2013

LearnBoost and Having Nice Things

Audrey Watters:

LearnBoost launched at the beginning of this recent resurgence in ed-tech entrepreneurialism, and in many ways, I thought it encapsulated much of the promise that new ed-tech startups are supposed to hold: great technology, great product, great team, grassroots adoption, freemium pricing, and so on.

To me, (co-founder and CEO Rafael) Corrales’ departure now serves to highlight some of the serious tensions, if not grave problems, that this new “ed-tech ecosystem” is facing. Indeed, what sort of “ed-tech ecosystem” are we really building here? Will it thrive? Which startups will survive? Whose values does this “ecosystem” reflect?

I admit I haven't paid any attention to LearnBoost, despite its being a competitor to SchoolTool (which I manage). Just looking at their website it isn't apparent how this is a business at all. I guess it is a "freemium" model, but usually there is something explaining what you might pay for. The product seems to be primarily aimed at individual teachers, not schools or districts.

Googling... ok, here's an explanation from Corrales:

Our intention is to keep our software free for teachers, parents, students, and even admins while having anyone who wants customization or extra support and services to pay a fraction of what the current systems charge. That's pretty powerful - we're going to save schools in the U.S. hundreds of millions of dollars a year, and globally we're aiming to save schools even more money.

Unlike our competitors in this space, we're well funded by the investors that backed Skype, LinkedIn, Twitter, and more. We're going to be around for a really long time.

OK, but that's a super low margin business. And that's based on my analysis as someone who is going to be pursuing the same strategy. I'm not sure if it will be enough to maintain the not-so-lavish lifestyle of me and one partner, let alone pay off for an investor.

And while common academic and data interoperability standards should be a boon to adoption in this business, won't standardization also, by definition, dramatically reduce the need for customization and support services, at least in the US?

It is hard for me not to conclude that the purpose of this company was always to develop some great technology and sell out to a big player. It doesn't make much sense otherwise.

Having said all that, I am a bit jealous of their technology. Timing is huge in a rapidly changing technological environment, and we ended up laying the foundations of SchoolTool right before everything changed in web development, and that's a cross we simply have to bear (or give up entirely).

Regarding open source, Watters praises LearnBoost and co-founder Guillermo Rauch for open sourcing lots of new infrastructure developed for their platform. And yes, that is a good thing. But also it is funny because while SchoolTool is 100% open source, one of the few hard rules Mark Shuttleworth gave for the use of his funding was to absolutely not hire web platform hackers and let them spend time writing generic infrastructure instead of application specific features. In general, it'll always be more in a geek's comfort zone to write, say, "the implementation of transport-based cross-browser/cross-device bi-directional communication layer for Socket.IO" than, say, a robust versioning system for academic standards or some other tricky business logic. It may work out fine for them, but it is a risky and expensive approach.

Also, open sourcing this stuff is admirable and helps the node.js community, but doesn't really do anything for the education community.

Regarding usability, if you're really writing enterprise software, the tradeoffs aren't so clear sometimes. For example, importing standards into LearnBoost takes one worksheet in a spreadsheet, whereas for SchoolTool it takes five. It probably takes two or three times as long in SchoolTool to get them set up. On the other hand, once you have that extra metadata, in there, you can automatically assign the correct standards to all the relevant courses in your school with one click -- saving you a lot more time in the end, but that's not obvious in a casual preview of the product. Our method is more abstract and probably does require more manual reading and maybe training, but it would save a ton of time over the years.

Regarding exporting your data, the big problem has been the lack of a standard format for doing so, particularly one that can handle complex school or district-wide data. Even if you want to dump the whole database in some easily reusable form, there's just not been clear method. Hopefully that's rapidly changing finally.

I don't want to end this on an "abandon all hope" note, because, oddly, inBloom may finally be breaking the ice to make the real solution politically viable. That is, big foundations just saying "Screw it, we'll just pay vendors with experience more or less successfully writing software used in schools to develop the open source infrastructure we want." Isn't that what Gates and Amplify did with inBloom? If that works, why not apply the same principle elsewhere?

No comments: