Wednesday, January 14, 2015

The Open Source SIS Market

I've been working on, essentially, a 10 year report on SchoolTool's development. This is one of the things which has been deterring me from blogging lately.

Anyhow, I did some retrospective research on the open source SIS "market," over the past decade, and it is somewhat of a cautionary tale for open source advocates.

Through the whole decade, there was a family of PHP open source SIS's in a more or less "complete" form, being used at some schools in America and elsewhere. We took a fairly brief look at the code early on, and it was pretty obviously terrible PHP. Like most early PHP, the code wasn't much more than a bunch of templates turning a database into web pages. If the templates are badly written, there is just not much to redeem the application. My snap judgement was that, if you wanted a good PHP SIS, to do anything other than start over from scratch would be a massive waste of time.

Instead, over the past decade, a succession of people have tried to redeem this codebase, forking the project in multiple directions, some investing non-trivial amounts of time and money in the process. I haven't followed these projects closely, spent any more time looking at how they work internally or externally. All I know is that none of them have become as popular as they should have. None of them turned into the Moodle of SIS's and seized a dominant position, even though they had every opportunity to. One reason I'm not naming names here is that I don't have any specific argument about quality other than something is wrong because they should have taken over six years ago but didn't.

This is a case of first mover advantage in open source, not just in terms of the application category, but language. Creating a new PHP SIS from scratch with superior architecture to the existing family would not be very hard. Gaining mindshare vs. the existing player is a big hurdle. "I'm going to solve the problems in the package you already know about" is an easier sell than "Move to this thing I've started from scratch, and by the way, I'm just some guy on the internet." Even starting a new SIS on a newer or perhaps more sophisticated language or platform is a clearer sell. This is an area (open source SIS) where the small profit isn't worth a lot of investment, marketing and advertising.

Essentially, there is a serious path dependency in a given product category based on the quality of the first mover. If latecomers are trying to redeem a faulty core instead of building on a solid foundation, the whole sector suffers.

One branch of this tree (at least) has become a successful commercial product.

One big challenger emerged from India: Fedena. It is based on Ruby on Rails and generally was built from scratch using modern web technologies. It might have swept the field except the prospect of making real money overcame whatever the initial rationale for open sourcing the project was, and they forked away and essentially abandoned the open source version two years ago from their ongoing commercial development. So that's that. As SIS's move more and more to cloud hosting, there's even less reason to try to market one as open source.

Shout out to Open Admin for Schools, a Perl-based open source SIS Les Richardson has maintained for schools in Alberta for probably 15 years or so, from which he probably makes some nice side income while saving schools in the region time and money.

Many, many people have written local SIS's and offered them as open source. This is a lovely idea but so far never works. It is just too much time to generalize, complete, package and market (even minimally) the product, particularly if all those tricky steps are seen as a side-light to a project which is probably the side-light of your actual job. They multiply the time involved by whole numbers, not fractions.

Finally, what about SchoolTool? Why didn't we take over 5 years ago, especially with relatively generous and consistent philanthropic backing? Well, I'll go into that more in the full report, and to be sure, I wish we had moved more quickly. But yes, we helped to clog up the market too. There were several years there where we seemed just around the corner from being "done" and having a "complete" SIS product, and with some influential backing, maybe we'd be a bad product to compete with.

Really just got to "complete" a year or two ago. And then growth can be very slow if you're talking school by school. You've only got one buying cycle a year, and people are waiting around to see if it works for other schools around them, so... it takes a while. We're starting to grow in earnest now. Hopefully it is not too late.

No comments: