Llima Orosa via Compfight

By Michael Horn, Columnist

“The technology is five years behind where it needs to be.”

It was the complaint of yet another school trying to build a blended-learning model that utilizes multiple providers.

“The software content providers are proprietary. It’s impossible to get data out of them. And when we do, the data doesn’t connect easily to the standards and the data from other providers.”

So went the grumbling from another blended-learning school.

What strikes me as most noteworthy about these comments, however, is just how un-noteworthy this state of the industry is in any industry.

At the outset of any industry, the technology tends to be immature and not yet good enough for the majority of users. In order to maximize the performance of the products and services and have any hope of them getting adopted, organizations need to integrate vertically and create interdependent architectures that tightly weave different components together to optimize performance, in terms of functionality and reliability.

As Clayton Christensen and Michael Raynor observe in The Innovator’s Solution, “by definition, these products are proprietary because each company will develop its own interdependent design to optimize performance in a different way.” The result of a proprietary, interdependent solution, however, is that customization is prohibitively expensive, because to customize, the company needs to re-architect the entire product.

But as an industry matures, the technology improves. It ceases to be “not good enough” for most users and begins to overshoot what most users need in terms of raw functionality and reliability.

As this happens, customers begin to prioritize new dimensions of performance. With functionality and reliability assured, they prize flexibility and customization, which proprietary products cannot supply.

The new solutions that arise to offer these customized solutions have a modular architecture — where different components fit and work together in well-understood and highly defined ways. Standards arise that specify the fit and function of all elements so completely that it doesn’t matter who makes the components or subsystems, as long as they meet the specifications. Modular architectures optimize flexibility, but because they require tight specification, it limits the freedom that engineers have to push the boundaries in terms of raw functionality.

These two states—interdependence versus modularity—exist on a continuum, but it seems to me that we may be at a crossroads right now in the blended-learning world between the two.

On the one hand, several blended-learning programs are continuing to use curriculum from one online provider, and although it doesn’t give them the customization they may prefer ideally, its simplicity and reliability are worth the tradeoff. Carpe Diem schools and the Flex Academies exemplify this–and neither seems to be complaining nearly as much about the technology.

On the other hand, increasing numbers of schools are adopting blended-learning models that have each student working with multiple software providers within one subject. But from their complaints, they appear to be pushing the industry toward modularity perhaps a bit before it is ready to shift and are therefore dealing with the corresponding headaches of a still immature technology.

At least one blended-learning school, Summit Public Schools, is partnering to build its own solution to the problem and use content from different sources to support the new competency-based learning model it is developing, which seems like a smart backward integration. Those demanding customized solutions seem to be running into headwinds. The fact that each of the schools has a unique model with different needs and requirements exacerbates the problem, as firm standards around which to coalesce just don’t exist yet.

Furthermore, some approaches to solving the problem seem unlikely to bear fruit. Standards are almost never negotiated among companies with proprietary architectures in an industry because the negotiations occur within a context where the representatives have the mindset of representing their proprietary architecture and trying not to get gored by the process. Much more likely it seems in the blended-learning world will be the emergence of a platform—like Khan Academy—on which lots of users write content that use the standards of the platform, as opposed to forcing a retro-fitting. The standards will emerge in de facto fashion, as schools vote with their feet—or clicks.

We’ll of course see how it ultimately plays out. For now though, the grumbling around the online-learning technology not being quite good enough is likely to be a refrain that we all ought to get used to hearing for at least a couple more years.

Michael Horn is Executive Director of Education at Innosight Institute, the co-author of “Disrupting Class.” He’s a graduate of Yale University and Harvard Business School. This post first appeared on Forbes.com