Updated: Apr 20
If you're already part of the Onoco journey, Margaret is likely a familiar name and face - after all, she is where the journey started!
Hello Margaret, tell us a bit about yourself!
Hello! I'm a proud mom of two and an entrepreneur - you might also know me as the founder of Onoco! A few years ago, I realized that there wasn't a single app that met my needs as a parent, so I decided to take matters into my own hands and create the ultimate app for parents like myself.
In my free time, I love surrounding myself with nature by caring for my ever-growing collection of over 20 plants in my flat, as well as being kept entertained by my two children of course. The marketing team have also convinced me to try my hand at being a TikTok star, so I’m having a great time sharing the funnier side of parenting on there
What was the motivation behind starting Onoco, and what goals did you have in mind at the outset?
I’ve worked for 10 years in product development in the Financial and Sports industry, building complex digital products, so when I became a mum for the first time I didn’t expect the product proposition from baby and parenting apps to be so fragmented and old fashioned.
I remember being frustrated that all of them were designed only for mums - I couldn’t share access to them with my husband and other caregivers. I decided to develop Onoco because society is changing and mothers are still carrying a huge mental load and are under the cultural pressure to be the fountain of knowledge of their families, but also dads are more involved in childcare and hiring a nanny to help is no longer stigmatised. Onoco, unlike most parenting apps, is designed to make shared care in a family easier.
What emerging technologies or trends do you see as having the biggest impact on the tech industry in the coming years, and how is your company positioning itself to take advantage of them?
I believe three of the most influential developments will be: decentralized technology for digital privacy and ownership, the growing adoption of wearables that are becoming smaller and smarter, and the expanding role of AI in supporting office work, creative work, and healthcare.
These three trends genuinely excite me, as I'm incredibly passionate about innovation in tech and how it can improve our lives. Recognizing the immense potential of these advancements, we recently launched Onoco AI, a cutting-edge tool that powers our Optimal Nap Time Predictor. One of the most exciting factors in this development is that it will continue to grow smarter as our research team expands and the latest multibillion-dollar AI models become open-source.
By combining various types of data, we aim to extend Onoco AI's capabilities beyond sleep and support parents in navigating the multitude of challenges they face. Considering that 90% of brain development occurs in the first five years of life, the potential for positively impacting families during this crucial period is truly enormous.
What inspired you to create the Onoco Nap Predictor, and what problem does it solve for users?
As a parent with a newborn, you're sleep-deprived and constantly receiving advice on how to put your baby to sleep when they're tired but not overtired. The question is, how can you tell? While hiring a sleep consultant is an option, many of them recommend following wake windows, which are often broad, non-personalized ranges that remain the same for both mornings and afternoons. This approach is unworkable for most families.That’s why we decided to build our own AI model and provide personalized wake windows and help parents determine the most likely time when their baby will be tired but not overtired, ultimately leading to a better night's sleep for everyone.
Can you walk us through the technical details of how Onoco AI works, and what data it uses?
In nutshell, Onoco AI is a neural network-based forecasting engine that predicts personalized wake windows for babies under one year old. The model uses anonymized data from over 1.5 million naps collected over two years, and incorporates data on sleep time, duration, baby's age, and other factors. We’ve chosen a multi-layer neural network in our model which was engineered over the past 12 months. It’s also worth highlighting that we protect user data privacy by anonymizing data and complying with GDPR regulations, never sharing any personally identifiable information with third parties.
For those who are interested in a bit more detail, I’ve described the whole process in our recent blog post available here.
How did you ensure that the Onoco Nap Predictor is accurate and reliable in its predictions?
As mentioned earlier, it took us over 12 months to develop this model. Its release timing wasn't due to the growing popularity of ChatGPT and the increased awareness of AI capabilities among mobile app users, although that certainly helped. More people now understand how powerful and beneficial such technology can be. Initially, we planned to release the model in 2022, but we explored various algorithms and models before deciding on the type of neural network we ultimately adopted. It was at that point when the model achieved 90% accuracy on the testing dataset, comprising nearly 200,000 test naps.
The next step involved conducting studies with real users. We ran three rounds of focus groups, providing access to the Optimal Nap Time Predictor for actual families using Onoco and soliciting their feedback. Based on the feedback, we either retrained the network or introduced UX changes in the app itself. In the final phase of testing, 83.3% of respondents indicated they would recommend Onoco Premium and the Optimal Nap Time Predictor tool to other families like theirs.
Can you talk about any challenges or obstacles you faced while building this AI feature, and how you overcame them?
The task of predicting a baby's sleep patterns is understandably a challenging task!
Perhaps the most challenging aspect for us was dealing with data incompleteness and dealing with outliers—such as when parents forget to log naps, data is inputted incorrectly, or records are maintained only for certain days (by a nanny, for example). Babies can also be noisy sleepers, leading to instances of extremely short wake windows (less than 5 minutes), likely when a parent rushes to the baby's room thinking they have woken up. We merged these types of naps together and treated them as one.
That's why we highly encourage our users to ensure they record all naps, as doing so increases the likelihood of receiving better predictions. We are refining this algorithm together, and it will become even more accurate for future users.
How does this AI feature fit into the broader technological roadmap for the company, and what other innovations can we expect to see in the future?
We aspire to lead the way in the parent tech space. Not only have we released our own custom AI model, but we also quickly adopt the latest technology introduced by Apple. Live Activities serve as an excellent example - this feature is part of the recent changes in the iOS16 operating system, which was first made available in December 2022. As soon as the initial documentation was available in the beta version, we began planning our implementation and timeline.
While Onoco's roadmap is not set in stone as we regularly respond to user feedback and innovations in the marketplace (as well as from our own team!), I can reveal that we are planning to launch a brand new section in the app focusing on innovative parenting content and evidence-based education. So keep an eye out!
What’s been your personal highlight so far since starting Onoco?
Last year, we went to a concert with my husband. It was a well known band and we were surrounded by a massive crowd. Just before the band started playing, I asked him about the arena's capacity and, when he answered, I realized that the number was equal to the families using Onoco that month. It was an overwhelming feeling. A few months later, we were featured as App of the Day on the App Store.
What movie or song describes your life at Onoco?
That’s a very tricky question! Sometimes it feels like Believer by Imagine Dragons…