Kinedu

Kinedu is a platform that gives parents or other caregivers a comprehensive, simple-to-use guide to activities and interactions that form the basis for brain-building moments. We focus on giving caregivers usable knowledge, accompanied with short videos of the activities, that can be translated to a change in behavior. Our solution would ensure that children’s learning has a timely start, and is guided by holistic, science-based guidelines. Through these interactions with caregivers, millions of children would get a better start in life.

Millions of children under five lose the opportunity to fully reach their potential every year. Simple, opportune interactions with caregivers form the basis for the development that these young brains need, yet millions of young children are under the care of adults who don’t know that this is a critical period for brain development or are not sure how to brain-build with everyday interactions.

In Mexico, one out of every five children ages 3 to 5 is behind on his or her motor, socioemotional, or linguistic development. This number doubles when it comes to the most impoverished families (UNICEF, 2018). Additionally, a large swath of the population is excluded from early education. Almost 40% preschool age children did not attend school due to issues with coverage.

Although the trend is strong in Mexico, there are similar gaps in early childhood education in Latin America, and many other countries of the world.

Kinedu is attempting to reach parents in order to provide the right information, at the right time. Mostly, we want to make it actionable, so that parents can become the main promoters of their child’s development.

Kinedu serves parents of young children, often first-timers, who want a guide to support their baby’s development. They often seek to know if their baby is on track, and want to know more about the process of development. A minority (≈34%) also want ideas for activities to entertain their child. A majority (60%) works full-time, and are looking to make the most out of their time with their child.

We conduct quarterly surveys with both freemium and paying customers to update our understanding of what users want from our app. Additionally, we carry out extensive interviewing with users every 8-12 months. We have honed our solution to 1) help users better understand their child’s development compared to others and 2) provide the option to download activities for internet-free browsing.

Through our interventions, we are also reaching low-income communities that might not have access to credit or debit cards (and therefore paid apps). Currently, we are working with 100 parents of children in vulnerable, informal communities within the metropolitan area of Monterrey. We designed a survey to understand their access to tech, and designed the intervention tailored to responses. We will hold interviews after interventions to improve the next round.

Kinedu is an app designed to help parents and caregivers engage in personalized, science-backed activities with their young child. Our app revolves around three main features. The first is a proprietary developmental assessment, based on World Health Organization and American Academy of Pediatrics milestones but honed and adjusted using data from millions of registered users. The second is a daily activity plan, tailored to each child’s development, which recommends simple yet powerful activities for caregivers to carry out with the child. Last is the ability to visualize and understand the progress that has been made by the child.

We have over 2000 activities, each accompanied by a short video that models behaviors (such as serve-and-return interactions) that are critical to the child’s learning. Activities work to strengthen the linguistic, cognitive, social-emotional, or physical realms.

Our app is available on iOS, Android, and web platforms. Content can be consumed in either English, Spanish, or Portuguese. Freemium users have access to a daily activity, while paying customers can access the entire catalog, as well as the full daily activity plan, composed of 4 activities.

We believe that we can positively impact a generation of children by building a sturdier foundation for future health, learning, and well-being. We believe that the key to this foundation is by empowering their own parents and caregivers to be the main promoters of their child’s development.

Kinedu is innovative in two dimensions: one, distribution, as we leverage mobile applications and their respective platforms to reach millions of users, and provide them with video guides that help them understand how to play with their babies. With mobile technology, we help parents who don’t have access to ECD centers in their community to understand how play can help develop, and act on that key driver of development.

Two, we leverage big data and machine learning in our business process but most importantly, our model of early childhood. Kinedu has processed millions of responses in our initial assessment about our users’ babies’ development, and with those responses, has built a model that describes early child development. We provide this model for parents through an assessment, and with the results from that assessment, we use an algorithm that personalizes the families’ experience within Kinedu. In order to refine and continue developing our model of early childhood development, we will continue to build our big data infrastructure and our machine learning models, leading to even better end-user personalization.

We leverage mobile applications and their respective platforms to reach millions of users, and use video to help caregivers understand how to play with their young children. We’ve developed our front end apps in each platform’s native programing language. iOS is built in Swift using their latest language, Android in Kotling, and Web front end in React. All Front-end solutions connect to our back-end through our APIs developed in Ruby on Rails. 

Big data and machine learning are also key to Kinedu. Since 2018, we have been using different Machine Learning algorithms to validate the psychometric properties of our Initial Assessment (Linear and Logistic Regresson, Dimensionality Reduction algorithms). Given of the complexity and non-linear nature of early childhood development, we are currently moving towards implementing Naive Bayes classification techniques, Random Forest trees to model and predict developmental trajectories and flows.

In order to refine and continue developing our model of early childhood development, we will continue to build our big data infrastructure and our machine learning models, leading to even better end-user personalization. Data warehouse is set up in Google Cloud, and our algorithms/models are custom built for our needs.

Our program strategies can be summarized as presenting easy to use, actionable information on ECD to caregivers. This includes sharing digested articles on ECD, offering information on the baby’s milestone accomplishment evolution, and providing developmentally-appropriate video-modeled activity recommendations.

Our targets are to increase the frequency with which the caregiver engages in developmental activities with the child, and increase serve-and-return interactions. We also seek to increase understanding of the child’s development, as well as knowledge on ECD/parenting. Lastly, we seek to increase his or her practical parenting skills.

As for the outcomes, we’re seeking to improve the child’s development in four developmental areas, strengthen the caregiver-child dyad, improve attitudes towards parental involvement, and increase parent self-efficacy. Possible moderators of these effects include tech savviness, initial parental attitudes towards involvement, and app usage.

To test parts of our TOC, we've worked with Prof. Frank at Stanford to look at how video-modeled activities change parental behavior. Results show that parents who are shown our videos increase the number of joint attention bids, as well as word tokens directed towards the child. We've also carried out small-scale tests in both the US and Mexico to test what effect Kinedu usage has on both parenting attitudes and children's development. Results show that more progress is seen in children whose parents show more Kinedu use, and there is a significant change in parenting attitudes towards the idea that learning at home in the first years of life matters.

We currently reach over 450,000 monthly active users. Over 3.3 million families in 180 countries have registered or used Kinedu in our 6 year history. We expect to be reach 1M active users in the next year, and 5M in the next five years.

Other than our customer base, we are also serving 100 families through our outreach and research projects in five of the most vulnerable communities within the metropolitan area of Monterrey. We expect this pilot to be replicated during the next school year in order to first reach the other 1,000 families who are part of the Cendi NL Vías No Formales program.

We are currently working to find other partners who can help us take these outreach projects to the next magnitude of scale, in order to keep testing Kinedu as an intervention tool in communities where access to early childhood education is scarce or nonexistent. In the past, we've carried out smaller pilots in Escobedo, Nuevo León, and the Roaring Fork Valley in Colorado, reaching approximately 60 families total.

Kinedu wants to become the leader in early childhood by providing tools and content for all adults with children in their lives. 

One of our biggest goals is to expand the suite of products we serve to other stakeholders and create an ecosystem around early childhood - most notably, teachers and caregivers. Our Model of Early Childhood Development, and ECD assessment, was created with the data from millions of parents - untrained adults interested in their babies’ development. This means that any adult - regardless of training, can have access to an understanding of development based on real world data. Our model accurately describes development: our algorithm should be able to recommend personalized, skills-based activity recommendations  for teachers, caregivers, and parents.

Generating this ecosystem will allow us to reach not only more children (as we drive a diverse group of stakeholders using our multiple, interconnected products), we will also provide a more consistent experience for babies in terms of their play/growth experience.

We’re also a private company with venture capital investors, and have our sights on constant revenue growth and profitability. Kinedu has grown 3x/year for the last 4 years and hope to continue scaling this fast. During the next five years, we should be able to reach profitability. We hope to reach tens of millions of users, in more than languages than those we currently cover - English, Spanish, Portuguese - by that stage and have a self-financing business for the future.

Our main challenges are the following:

Scaling - efficient customer acquisition - is our biggest challenge for growth. Kinedu needs to develop alternative acquisition channels to our current solutions (paid social media and organic, off-line referrals). The more efficient we get, the easier the rest of the barriers get.

In terms of financing, while we currently have significant cash in the bank, but will need additional rounds of funding to keep the pace of growth.

In order to reach many more users, our content and apps will need to be available in multiple languages (currently, English, Spanish, Portuguese), and localizing our content will need new team-members in different regions who are adept at local, cultural contexts that we need to cover.

We are doing a lot to improve our efficiency scaling and lowering customer acquisition, particularly around strategies which leverage additional channels - working on SEO; working on improving our blog (which is free content online; and designing and building a solution for additional stakeholders that can bring parents into the Kinedu ecosystem. This is in addition to optimizing what is currently working - Facebook, Google, and the experience within our freemium tier.

In terms of financing, we are constantly working with potential funds to understand our fit within their portfolio, even when we’re not fundraising.

FInally, in terms of languages - Kinedu recently expanded into Portuguese (July 2018) and have learned a lot from that experience. We recognized the need for a local lead with knowledge of parents and local culture; We recognized the need to give adequate time to localization and contextualization. However, we also built the technology to “plug and play” content once it’s localized; and we used machine learning to translate 95% before revising with a local editor. With Portuguese, we understood the challenges as well as our advantages to the expansion.

Our team includes 42 full-time and 2 part-time team members in its Monterrey office. We also have 2 team members based out of Santos, Brasil.

Our team combines business know-how, tech prowess, and a deep commitment to and domain knowledge of early childhood development. Luis Garza, founder and CEO, previously founded Advenio, corporate daycare centers now in Monterrey, Mexico City, and Bogotá. Luis has a degree in Management Science and Engineering from Stanford University, and an MBA from Harvard Business School.

Our Content and Product Director leads a team of in-house psychologists focused on early childhood development. Our CMO has years of experience understanding our end customer: parents of young children. Our tech team has specialized teams for Android, iOS, and Web development. Our Analytics team ensures that our data is driving continuous improvement of our product. Our Science and Special Projects team is dedicated to testing our Theory of Change, as well as managing our research initiatives with Stanford’s Michael Frank and our participation in the Center on the Developing Child’s social enterprise initiatives.

Kinedu has a unique focus on data-driven, science-based recommendations for parents. Our team ensures the sustainability of the model while ensuring that our offerings are evidence-based, and really having an impact on the children and families that use our products. Last but not least, with more than 3 million families in our total user base, Kinedu has the largest known database of early childhood development. This allows us to continuously improve our Developmental Assessment (with input from Michael Frank’s Language and Cognition Lab), and contribute to the science of early childhood development.

We work with Prof. Michael Frank and his Language and Cognition Lab at Stanford University to 1) use data models to improve our Developmental Assessment, incorporating the data from our back-end to newer iterations of the model and 2) test specific parts of our Theory of Change. For example, Prof. Frank’s team is measuring the impact watching a Kinedu video has on joint attention between a child and his or her caregiver.

We also have established research partnerships with the municipality of Escobedo, the Valley Settlement Project in Colorado, and CENDI Nuevo León to research the impact Kinedu can have when used as an intervention tool in low-income communities. With CENDI, we are currently measuring impact in five different communities in the Monterrey metropolitan area, using Kinedu to promote developmental activities over summer breaks.

We also have partnerships with advocacy and ECD organizations. As an example, Kinedu is a funder and founding member of the Aceleradora de Innovación para la Primera Infancia, a Frontiers of Innovation cluster for Mexico. We also participate in the Center on the Developing Child at Harvard University’s first social enterprises cohort. Our content has been featured in the Center’s communications on serve-and-return interactions.

Kinedu provides our content in a Freemium to Premium subscription model for parents; we acquire users through different online channels (FB, blog, Google search, etc) and let them fall in love with the experience of our freemium tier, which get a free activity a day, plus basic developmental progress reports. We incentivize purchase through promotions and others. Once premium, users can get 5 personalized recommendations per day, unlimited access to the catalog, in depth progress reports, up to 5 babies in their accounts, and unlimited caregivers (other adults). Through this, we help parents understand how their babies’ are doing and what they can do to encourage development. The end users, and main beneficiaries, are their kids, who receive more attuned playtime during their early years.

Kinedu will continue to scale revenues as we acquire more families to our paid tier. We only have 30,000 paying subscribers and believe we can scale that substantially. Also, the massive amount of users lends itself to other revenue models such as ad-supported; affiliates, amongst other. We will need to raise additional rounds, but at this point have no immediate intent.

Kinedu would greatly benefit from the 12 months of personalized support from Solve. For example, we are currently developing Kinedu Classrooms - a way to reach caregivers in center-based settings with the curated content that we have already developed and tested with our millions of users. We realize that mentorship and connections through Solve would greatly increase our chances at successfully reaching the populations that would most greatly benefit from our products. The MIT and Kendall Square community also has so much to offer in terms of helping us build out our Machine Learning models.

Finally, part of our core beliefs at Kinedu is that our work should be evidence-based. Currently, we are working on our largest-scale intervention in five low-income communities in the metropolitan area of Monterrey. The Solve community could help us direct our efforts to ensure that they are being spent in the right direction.

We believe AI and big data can help us improve our model of early childhood development skills as we gain more info from families across the world. The model itself is exciting from a research point of view, and we see an even bigger opportunity in refining the personalization engine (currently a linear algorithm) to provide activities that better attune to the individual child’s level, as well as the parents’ skills and affinities. We would use to prize to help set up up our big data infrastructure through a data warehouse, and continue refining our model for recommendations to parents that have a real, measurable impact on their child's development.

Kinedu has a great opportunity to be leveraged in the United States, our largest market. The high cost of quality daycare, and the lack of access in many communities, will likely be a key driver of inequality in the future. However, Kinedu provides cost-effective tools for any adult with a child in his or her life to contribute to the important brain-building taking place during this key period of development. Kinedu is also developing Kinedu Classrooms, a platform that provides caregivers in group or center-based settings with the right tools to provide developmentally-appropriate activities tailored to the needs of the individuals in the group.

If selected for the Innospark Ventures Prize, Kinedu would use the prize to help fund our hiring in our Analytics department, with a focus on finding individuals that could help refine our model for recommendations to parents. We would also fund an intern program to identify and help train Universidad de Monterrey students with an interest in using data to drive an understanding in development and adult learning. Finally, the funds would help set up up our big data infrastructure through a data warehouse.