🚀 How to grow your startup with a great product loved by users

It has been a year since we started, Cherni, my cofounder, and friend, and I to give shape to our ideas, hypothesis, and vision.

In a first post, I wrote about our product design process, and how to use product design and data analysis techniques to maximize our learning during. You can read it here.

In this post, I want to explain in more detail our research approach, and how we validated the hypothesis that we have frame along with our journey, and how we interview potential users and customers.

The Problem

We started with the scientific fact that physical environment, habits and social factors account for 60% of our health.

This means that someone is not healthier if it has the best health insurance or has the best hospital around the corner, they help, but they are not significant to keep us healthy.

Another huge problem is the current demographic pressure over health systems in developed countries due to the lack of resources or the lack of health services in undeveloped and in developing countries.

Plus, other factors such climate change, unhealthy behaviors and the risk of huge pandemics, all these factors are a threat to our health.

Therefore, along with our advisors, we draw a conclusion, which is that in the future a person will only have secured its health if she has the money or starts to have healthy habits and prevent environmental risks since young ages.

From that, we had huge vision, and we still have, to help millions of people to have healthier lives and build digital health insurance company on top of preventive health platform, that is capable to measure and predict physical environment, habits and social factors of a person and how all these factors affect our health.

Now, all of these is possible, we have tons of digital breadcrumbs from our behaviors and social interactions, and we also have tons of data from our environment. Using all that data with machine learning models we can build personalize models for each person to predict the next symptomatological state, based on environment, habits and social interactions.

For example, we can predict that someone is going to have a stressful or a headache the next day, because there are high levels of pollution, along with high levels of pollen, and there is high amplitude on temperatures, and so forth.

Knowing that high levels of stress over a period of time can cause a depression of the immune system, and therefore a higher probability of the environment and behaviors to affect our health.

From this starting point we started small and build two prototypes based on public health alerts, such as airborne diseases, pollution, pollen, extreme weather, lice, mosquitoes and many more, and we tested if people are interested in them, first with Facebook Messenger and later with a mobile app, after analysing our data, we realised that we had user retentions of weekly 40% after 12 weeks, with nearly 1K users that we were capable to persuade.

However, we realized that we needed a more powerful solution to scale, therefore, we started over from scratch, and we came to a clear starting point.
Respiratory patients with asthma, COPD or allergic rhinitis need to know the triggers from their respiratory crisis, and mostly they are due to environment, habits and social factors.

From this new hypothesis, which seems reasonable to start small, focus on a very clear niche and build our technology, we started our research, which explains further now.

How big is the market? and show me the money

There nearly 800 millions of people in the world with asthma, COPD, and allergic rhinitis. These means that millions of people have daily difficulties to breath.

Ok, but we are in Spain, as a playground market for our hypothesis and technology, it means that we have a home market of more than 5 millions people with respiratory patients.

We know that Spain maybe is not the best country to start a company and even more difficult to start a health company. Due to cultural aspects. However, we thought that we success here, our solution will be replicable in any country.

How does it look the health market in Spain? In the following drawing, you can see a network of cash flows, and who really pays the party.

This strategic task helps us to understand that we first need to build a great product for patients, and then go to health insurance companies and pharma companies.

So that is what we did, we went out to talk with respiratory patients, neumologists and allergologists, health insurance companies, and pharma companies.

Finding vitamins and painkillers

Normally, innovative products start being a vitamin and if the execution and the adoption is significant they will become painkillers.

Our digital health diary without no doubt is a vitamin, we, as humanity, have lived millions of years without a data drive health diary, but now is possible.

The same happen for Facebook, Airbnb, Google and so on, but now they are totally incorporated in our daily behaviours.

So, we need to iterate and learn as much and as fast as possible in order to build a painkiller that will help millions of people, we some ideas on how to do that.

If you want to read more about vitamins and painkillers, this is a great article about it.

So which were the steps that we took to give shape to our product?

Online surveys

First, we used online surveys to test major ideas, behaviors, concepts and pain points in respiratory patients. We ran two surveys one focused on health behaviors and another one on asthma patients lifestyle. They helped us to have the first view of our users.
For the task of doing survey analysis, I recommend Graphext, it is a great tool for data analysis, that help us to have insights very quick, without coding a line of R.

We survey more than 250 people about health habits and nearly 50 people with respiratory problems. We learned about which are their interests, pains, and traits.

These helped us to build a clear persona which will use our product.

Interviews

After that, we worked so hard to interview all stakeholders, patients, doctors, health insurance companies, and pharma industries.

We interviewed 15 people with asthma or asthma carers, these interviews helped us to understand first hand how they live they daily life, how to treat their health conditions, what problems they have to stick to their treatments or keep track to their symptoms.

With doctors we talked about what is the general practice to treat respiratory diseases, such as asthma, COPD, and allergic rhinitis, we interviewed four doctors, and some of the interviews were very insightful and helpful, we found a clear problem in the practice, they fighting to increase the adherence to treatments, however, users are not, and we found that treatment adherence is a behavioural problem.

With all this information we build a story to go to talk with health insurance companies, and we found that our story resonate, they spend tons of money on respiratory diseases, and despite to other chronic diseases, in three months patients can be under control and with better quality of life if the follow their treatments.

We made contact with some pharma companies, however, we didn’t have any contact with the industry and was much more difficult to get real conversations, and not over an email.

Jobs to be done

So, because our main product right now is focused on patients, we designed our product around them, based on our learnings, and trying to reduce the friction as much as possible.

Some of the tasks that we identified that users should be able to do are the followings:

When there is something that can cause me asthma, I want to know it, so I can take my med and go on with my life.

When I have been a few days with symptoms, I want to know if I’m under control, so if not I can make an appointment with my doctor to adapt my treatment.

When I have respiratory symptoms, I want to know if they are caused by the environment, so I can avoid those triggers.

Here, you can find a great article about how to design Jobs to be done, from my point of view a think is a great framework to design products, much more than designing personas.

Prototype and iterate

We prototype multiple versions of our product, always, using mockups, here you can find some of the mockups we built. And we test nearly daily with users. To find what is clear and not, and more important to learn as much as possible from them in order to persuade them to use our product.

Here, I can only say one thing, be open to feedback, and when I mean open, you should be capable to accept hard and uncomfortable feedback, if you only get good feedback, that’s bad.

User tests

In any iteration we tested several users, I can remember how many, but I’m sure that more than 30 people, friends, colleagues, family, even random people at Google Campus Madrid cafeteria.

Here, you can find some tips on how to do user testing, from the great Marc Borras.

Product meetings

Every month we had from one to two product meetings with our product advisor and friend, Carlos Sánchez.

He made sure that we kept learning and we don’t bump into any distracted task and focus on solving a real problem.

These are Cherni and I trying to reach somewhere, and Carlos, taking care of us.

I might say that I have learned so much from his feedback and advice on how to build products. I’m an electrical engineer with a very strong background in mathematical models and data science.

I haven’t been taught to solve problems for people.

Behavioural analysis from previous product iteration

I talked about retention, so I got data from our ongoing running prototypes, and these are some of the results we got.

This analysis helped us to build a better understanding of user minds and design an onboarding process to maximise engagement in day 0, which will help us to improve our current retention base line in the future.

First impact is what it counts. So we need to focus the coming weeks on day 0 and 1 in order to delight our users.

For behavioural analytics I recommend this playbook from Amplitude. During the process of product development, we invest a significant of time in defining an events taxonomy, integrating them into our product, and our north starts metrics.

I also encourage you to read this post from Brain Balfour about metrics and how to explain and predict them. On this subject, I have a clear view, which is to build a very solid analytics platform for explaining and predicting capabilities. In the meantime, some raw data and some R models will do the work.

Some learnings

Reduce the friction on your product by increasing the ability of the job to be done by the user, associate that job with internal and external triggers, and motivate your user to carry out the job.

I always thought that designing and building products were going to be easy and fun, I’m a frustrated architect, and I always love the design and thought that I have good taste, but I have realized that it’s fun but very hard and stressful.

Are we building the right solution? Is it going to work? Will our users love our product?

I recommend this article about the 10 principles of good design from one of the most famous product designers of all time. Dieter Rams. I’m sure that many of you had at your home some of the designs or products inspired by them.

When launching your prototype try to build network effects, but start from small-medium nodes with high centrality, they will bust your growth and they will give you negotiating power. This an advice from my learning on social computing. Virality starts small.

When I build products, I always think about future features that data will allow me to do, give me data I will do magic.

However, when you are building an innovative product normally you don’t have such data, you have to collect it to improve your product.

This means that you need to give value from day 0, so if you don’t have that data that will allow you to build very fancy machine learning models, try to solve a problem from scratch, and collect data since day 0. If you know how to use it your product will be loved.

And if you don’t have such data, find a proxy to build such features. As always start small and do incremental learning and improvements.

If you are building data products, you will have faced the chicken and egg problem, which is related to the data gravity concept. Here you can read more about it.

What now?

I’m waiting to go deep on behavioral analytics, build fancy statistical models to understand our users’ behaviors and what makes them love our product.

At some point, I will write about our experimentation framework and how it works. the next three months we will be focused on test it and improve our metrics to raise some money.

The main objective right now we need to build an audience of users who love our product, by listening to them and build what they want.

Here you can see the result of our product design process, a digital health diary for respiratory patients, which we want it to be the best and smartest health app, upon the latest research on social computing, digital epidemiology, and machine learning.

👇 You can download it here. Sorry, only Spanish version right now.

📱 iOS version

📱 Android version

I want to thank all the people that make the effort to make some time to answer any of our surveys, interviews questions or give us feedback, and special thanks to Carlos Sánchez, Lucía Salem, our advisors, friends and family for supporting us.

👼 We need your help to spread the word, share it if you like this article or share the digital health diary for respiratory patients with your family or friends groups in Whatsapp if you live in Spain.

Further readings:

5 Habits to build better products faster from Hiten Shah

How I grow mint.com from zero to 1 million users from Noha Kagan

David Martín-Corral

Written by

I build data products 📈. Co-founder @zensei_app and @politibot. I research on complex systems, social computing, ML & digital epidemiology 🤒👩‍👧‍👧 @uc3m.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade