The global edtech market generated $17.7 billion in revenue in 2017 and is expected to grow to $40.9 billion by 2020, but bad design means much of this investment won't pay off in the classroom.

A selection of renowned edtech researchers and academics at the Global Education Skills Forum explained why this happens and gave their tips on how startups can design tools that work.

carla aerts
© Tom Atkinson

Don't retrofit

The root of many of their problems is in a system that doesn't easily integrate innovation, according to Carla Aerts, the director of Futures at UCL Institute of Education, known for her research and trendy glasses.

She says that edtech developers typically try to retrofit their products into the education system or mimic a teaching process.

This doesn’t work and in fact can cause more harm than good, by adding to a teacher's workload and shifting their focus from their students to a screen.

"What does the best technology do?" asks Aerts. "It transforms, it doesn’t retrofit."

She says that rather than look at what teachers do, edtech must look at what they need.

Design for engagement

The correlation between engagement and learning that teachers observe every day has been scientifically proven.

Philipp Schmidt, the Director of Learning Innovation at the MIT Media Lab, points to research at his university that revealed how it works by monitoring electrodermal activity of students while they worked, rested and played.

A wristband tracked the electrical conductivity of their skin to understand their emotional responses in different situations over one week at MIT. The lowest level of engagement was observed while the students sat in lectures.

This could transform the way we look at classes, but such systemic change is beyond the reach of most startups. Schmidt believes that a better option for edtech designers is tweaking current practices to boost engagement by making learning more active.

Think small

The most clever and advanced edtech is often driven by engineering and technology rather than by pedagogy. The creators of platforms tend to think big when defining their product, and end up with something that only works in one context. Education systems vary immensely between nations, regions and institutions and can’t normally be generalised.

Instead of solving the world of education, Aerts recommends startups to focus on one specific facet. They can find this by learning from the stakeholders, whether they’re learners, teachers, parents, or school administrators.

They can then identify their problem and design a solution to resolve it.

If they focus on a single component, they also can more easily draw on prior studies and identify how others have overcome similar hurdles. They can also validate the idea through testing it in a precise market.

"Identify some key regions or markets for you and then validate it in those markets," advise Aerts.

"It also gives you the flexibility to improve on what you’re doing or where you got it wrong. If you're going to fail, fail fast, fail smart and start again. Don't build something big then come crashing down and [then] have no way to move in another direction.”

People learn through conflict

Swiss psychologist Jean Piaget's pioneering work in child development led him to understand that cognitive conflict is what leads to learning.

When our beliefs are challenged in a way that forces us to engage we learn from the dispute.

"Learners will argue, and in that argument is the seed of learning," says Justine Cassell, an associate dean at the School of Computer Science of Carnegie Mellon University and co-chair of the Global Future Council on the Future of Computing at the World Economic Forum.

Her lab uses machine learning to predict what piques the curiosity of students. The results show that it grows through internal cognitive conflict, but is far more likely to arise when a group of learners argues among each other. The curiosity this provokes leads them to explore different ideas and learn new things.

Social environments support learning

Technology that creates a social context for learning consistently proves to be more effective.

"If there's one thing that you need to know, it’s that presenting information is not going to make learners learn," says Cassell. "It's not going to help them. Learning is fundamentally social."

Schmidt refers to an experiment that studied how young children learn languages through sounds. It showed that they learned the most when they had a meaningful relationship with person who was speaking the language to them. Their brains paid more attention to this person than to an audio recording. The feeling of rapport that improved their learning can be digitally developed.

Schmidt noticed while working with a student in a group called Personal Robots, which designs robotic learning companions that work alongside educators. They found that they could build trust for the robot by using micro social cues, such as nodding at the right time or providing encouragement during a nervous pause.

 “Think meaningfully about how you can engage other people and how you can engage the technology as a social partner in learning,” suggests Cassell.

Choose investors who care

Another challenge for edtech startups is that they need funding to develop and investors often want a short-term return. Reliable evidence of edtech efficacy will only emerge in the long-term, if it emerges at all.

Success will vary among institutions, ages, regions and values and is hard to quantify. The research question often isn’t clear and any results will contextual.

Researchers can demonstrate that their product improved test scores by a percentage, but will struggle to isolate its impact from the other variables. It's also hard to prove that the solution is scalable and replicable in different contexts.

Edtech startups must therefore pick their investors carefully.

"If you get an investor who’s got passion for education, they know it's not going to generate a big pot of money in a very short time," says Aerts. "They probably know it’s not going to generate a big pot of money in the first place."

Focus on user experience

Edtech tools will always fail if the no one wants to use them. Teachers have to deal with numerous students with varied and often complex needs. Picking up a device can break their connection and push pupils to the periphery of their work.

The technology should streamline their workload and support the connection with their students.

Students often have greater digital acumen than their teachers, but also have higher expectations. Interfaces determine initial engagement, but a lot of edtech is chunky and far below what students are used to in video games and films.

Aerts recalls working with a company behind a virtual world designed to help learners prepare for their GCSEs and A Levels. The company trialled the product with a group of GCSE students, who were distinctly unimpressed. The superfluous aspects of the platform were a time-wasting distraction from their learning.

"Don't come up with a lot of paraphernalia, because a learning process or a teaching process is incredibly focused, and what you want to do is you want to augment and enable that to happen," says Aerts.

"You don't want to detract by bells and whistles. What you want to do is really enable this to happen, and that can only happen through good and engaging design."