Edmodo

Edmodo connects teachers and students to a world of other people and resources. Learn more. Teach better.

Teachers and students are both better off working with like-minded peers. Students have 3x better completion rates when studying in groups and 2-5x better learning outcomes. Edmodo makes it possible to find and work with peers anywhere in the world. Users ask questions, others suggest answers and robust discussions ensue. Resources are often shared: in 2017, over 110M unique resources were shared by teachers.

At Edmodo, our goal is to facilitate communication and collaboration, and use the sharing activity as a curation signal to find grade appropriate and need-relevant resources for all users. Resources are recommended to users based on the communication patterns observed on the page.

Teachers and students have profiles that are updated automatically with resources shared and consumed. Since all activities occur on the platform we can also trace the impact of any given individual over the years: what students were impacted by a teacher, due to which resources, and how much.

We will identify the best teachers and resources for all subjects at all grade levels anywhere in the world.

There are two basic hypotheses behind Edmodo: (i) teachers and students talking to other students, asking and answering questions, exchanging pointers and recommending resources leads to better outcomes, and (ii) machine learning can extract patterns, signals and profiles from these activities to recommend connections, resources and behaviours to users just in time.

To this end we have built a communication network for teachers and students, examined patterns of connections and interactions, and started recommending resources to other users in similar situations. Students and teachers can benefit from  tens of millions of other users and their experiences and outcomes.

The communication and collaboration platform is built using stream based network models first popularized by Facebook but built internally using a Go based backend with a React based front end. Our iOS and Android apps are natively implemented.

The machine learning platform for recommendations uses a word-embedding model loosely based on the popular word2vec papers but modified to take into account our platform idiosyncracies and limitations.

We power more teachers and students than any other platform currently in place, all due to technology.

1. Internationalize the platform
2. Better recommendations with distinct models for teachers (who tend to be more task oriented and focused on a goal) and students (who often tend to be more browsing oriented and less focused)
3. Better profiles for teachers and students and move away from explicit assessments to a model of more implicit, continuous model building

1. More connections to students and teachers will improve recommendations to users: we will need to make the advantages of this network clearer to our users through signals in the product
2. Better and larger lists of resources used by teachers in schools around the world (we currently know of 650M resources recommended by teachers to classes over the last 8 years) and their efficacy for different users
3. Better user and resource models and relevance algorithms, with continuous adaptation of the two for real time matching.

Word of mouth through existing teacher networks has been our model of growth so far.

We currently have about 85M registered students, 8M registered teachers and 3M registered parents. Most of our usage is in schools during the weekday and at home during weekends.

Our growth has slowed, so it is difficult to predict, but we hope to grow to about 100M users in a year and 120M users in the next 3 years.

We have a very experienced senior team: two of the executives founded an educational technology startup that they took public in 2004; another exec founded a security startup that he took public in 2004; yet another was an early Google researcher; our original investor team included early execs in Facebook and LinkedIn.

Short term revenue will be through ads and courses for teachers. Longer term we plan to serve the resource and service needs for students and parents as well. We believe that this is a great area to be in.

We are applying to SOLVE because education is too large a problem to be left to any single group or organization. It requires the collective thoughts and ideas from everyone involved. SOLVE is a terrific resource for us to try and work with, both in terms of ideas they have, but also in terms of their network and connections.

• Need to understand best practices in education, different standards around the world and how expectations are changing rapidly in response to connectivity and mobile availability
• Technological changes in AR/VR and machine learning and how these may be used in the classroom and at home
• SOLVE has access to the world's greatest experts in education and technology