Recent investments pouring into digital language-learning have once more uncovered significant developments within our industry. Traditional textbook publishers, however, seem to constantly close their eyes for these developments. Will they survive the value innovation created by language-learning EdTech companies?
Digital vs. Algorithm
While traditional textbook publishers remain focussed on their print-first strategy with elements of digitalisation, new ventures are rapidly gaining ground with their algorithm-first strategy. Where the print-first strategy merely explores possibilities of distributing existing textbook content digitally, the algorithm-first strategy relies on automated creation of that very content, along with highly adaptive delivery mechanisms. Algorithms instead of human authoring, rendering a high-volume, low cost business model.
New kids on the block are building automatised generators of learning content based around machine learning, computational linguistics and artificial intelligence. What they are striving towards is a highly cost-efficient content generator that adapts maximally to learner’s needs. The latest investment into these kind of companies only confirm this development.
- A great example is Duolingo that recently received 45M in series D investment led by Google Capital. Duolingo mostly focuses on gamified language learning with automatic task creation.
- Another great example that just cashed some 8 million is Lingvist - an Estonian startup focused on learning speed and efficiency with great support of algorithms.
- On the other side of this spectrum are dedicated digital publishers of learning content such as Newsela (that raised 15M last month from among others Mark Zuckerberg). They start with the high-quality manual approach and slowly move towards automatisation of the publishing process with help of the same mix of AI and computational linguistics.
Again, what all of them have in common is the large automatisation of content production, adaptability to learner's needs, and clear progress monitoring.
It is hard to believe that content, that needs to be authored and edited manually, can continue to compete with highly automatised and adaptive content providers.
Traditional publishers still rely primarily on static content with multi-media enhancements (essentially a book made available in digital format). It is hard to believe that this static, one-size-fits all content, that still needs to be authored and edited manually, can continue to compete with highly automatised and adaptive content providers in the near future. The developments in the EdTech industry have the words ‘value innovation’ written all over it: equivalent or better service (language-learning), but at dramatically lower costs.
Naturally, language learning is not only about technology where large amounts of content is pushed to the learner at lowest permissible cost. A big part of the learning process still happens in the classroom where learners write, speak and go over grammar. This process has been traditionally supported by the static content provided by educational publishers. However, we are rapidly heading towards a situation where the digital-first providers not only offer content, but also provide pedagogically sound, engaging and affordable tools that can be used by teachers and students anywhere, anytime and on any device. Think big data and learning analytics.
Delivering static content in an online environment was an important first step but future language learners and teachers will have access to affordable, highly personalised and engaging language learning technology. Educational publishers should practice what they preach and learn from computational linguistics and artificial intelligence. Maybe not to survive, but to stay relevant at least.