Note: Since the publishing of this article, the Philips NutriU app has grown in scope and changed its name to HomeID. You can learn more about the project by visiting our work page.

Many years had passed since the Fourth Industrial Revolution knocked on our door and among other things, brought IoT devices to our doorstep. The application of these devices is unbounded – the technologies are already used in so many areas of our lives and are still growing and evolving at an unprecedented rate.

Humanity came to the point where there are more connected devices than living people on this planet. With so many different apps and products on the market, users have the freedom to choose what fits them best. They will not hesitate to choose a product from a different brand for a better experience, so understanding their behavior is critical.

Since more and more brands are competing for users of smart domestic appliances, we wanted to really make the most of the data available for our client Philips. This means all the data, from multiple sources. We combined telemetry data from the smart air fryer with user behavior data from the NutriU mobile app and got great insights for improving the user experience.

Together with Philips, we are creating a content-rich, globally available mobile app that helps users find inspiration for cooking, connects to their smart appliance, and allows them to control it remotely. For example, this means users can control their smart air fryer from the sofa while watching their favorite TV show. They can extend cooking time, lower the temperature, see how much time is remaining, or just put it in warm mode to keep the meal warm until the end of the episode. 

Both cooking beginners and pros can easily find recipes to fall in love with. The NutriU app, which is available on both the App Store and Google Play, opens up cooking horizons and provides a new way of healthy cooking that saves time and improves cooking skills.

For getting to know our mobile app users, we already had Google Analytics for Firebase in place. We defined many data points to track each of our feature’s usage.

All these data points are documented in the so-called event tracking plan. With that data, we can answer many questions that are important for app growth. These are just some of the things we wanted to understand:

For this specific case, we wanted to connect all of that mobile user behavior data with the data from the appliance itself. The latter reveals which recipes and presets (predefined time and temperature that users can select easily) they use, or if they select their own time and temperature, and most importantly, what they change in the process.

All these data points are collected with Amazon Kinesis Data Stream which is a proven tool for collecting streaming data. The raw data is first stored in the Amazon S3 bucket and then sent to Google BigQuery where we store our mobile app data.

We needed to understand which cooking data points from the appliance belong to which mobile app users, so we assigned a unique device ID on both sides. Each device sends the data together with its unique device ID and each mobile app user is assigned the same ID after successfully pairing the device with the mobile app. 

Next, we transform these cooking logs into meaningful sessions and aggregate them to understand the details of each cooking session. Each session is connected to the mobile app customer and their smart device, limited to the users who gave us analytics consent, of course.  This enables us to understand how they take advantage of the mobile app to enhance their cooking experience. 

All this data is updated on a daily basis and integrated into the Google Data Studio dashboard where it is visualized with compelling charts. The dashboard enables us to truly understand user behavior across apps and smart appliances and use it to make their cooking experience even better. For any deep dives into the data, we always have easy access to the database.

This data enabled us to make data-driven decisions based on user behavior. By combining all the sources, we were able to utilize the data to support business objectives. For example, we increased the connection success rate in the mobile app’s onboarding flow by identifying the pain points in the onboarding steps. Making these changes based on actual user behavior helped us improve the user experience and retain our users.

When you have your customer data available, don’t let it just sit in the dust. Make something out of it. If you don’t have any customer data, well, it’s about time to start collecting it. For some products, analyzing just the mobile data won’t be enough and it will be important to understand the users from a different point of view. 

Companies need a 360-degree view of a customer to understand their needs. It means having all the data from many different sources aggregated in one place. That can be user behavior data from a mobile app, IoT data, purchase data, and much more.

If you feel unsure about any of the steps in data analysis, feel free to reach out to us and we’ll jump into your data head-on to grow your product.

And if you’re interested in other ways of gaining a competitive advantage in the Internet of Things realm, check out our IoT solutions page.

We live in a time when 3.5 billion people are using smartphones – that’s almost 50% percent of the world population. The enormous number has led to a staggering 204 billion app downloads since last year, and brought in more than 120 billion dollars revenue.

In such a competitive market, all players are trying to reach as many users as possible, engage them continuously, and retain them for as long as possible. The key to success is knowing what app users want and using this information to improve the apps.

In-app event data is the source of that information for you.

From the moment the user installs the app, their actions can be monitored and recorded as in-app event data, such as the record of registration, login, or in-app purchase.

By discovering users’ actions, it’s possible to make data-driven decisions that can help bring more users to the app or retain the existing ones and make them happier.

To make sure that we have a quality and consistent stream of data being collected from an app, we need to make sure to write an excellent event-tracking plan which will be the cornerstone of our data collection.

However, the first step in creating a solid event-tracking plan is determining where to send the in-app data.

According to the research from May 2020, out of all Android apps that use analytics SDKs, 95% have integrated Firebase, as opposed to 40% of iOS apps.

It’s a powerful analytics tool that measures everything from user engagement to app crashes. It’s also practical for quick tracking setup, as some events like screen views and session start are tracked by default as soon as the SDK is implemented in the app code.

Besides automatically collected events, Firebase offers great features such as Cloud messaging, In-App Messaging, and easy-to-use Crashlytics.

One of the most useful and regularly used features at Infinum is its integration with Google BigQuery. It’s a serverless, highly scalable, and cost-effective data warehouse designed to help to turn big data into information.

Also, having data in BigQuery gives the possibility to integrate it with almost any popular visualization tools like Tableau, Qlik, Sisense, or Google Data Studio.

In the words of HP’s former CEO Carly Fiorina, “The goal is to turn data into information, and information into insight.” Having access to raw data is a must because it isn’t possible to make a deeper analysis in every mobile analytics tool.

When it comes to building event-tracking plans, here are some of the most used analytics terms to be aware of:

Event – User’s each interaction within the mobile app.

Example: A song played, user invited

Event property – A parameter containing additional information about the event sent along with that event.

Example: If an event triggers when the user plays the song inside the app, song name and artist name can be sent as event properties

User property – A parameter that can give information about each user individually.

Example: User that opted in for analytics or user that opted out for analytics, users using the app for free or premium users

Screen – The name of each screen that exists in the mobile app.

When it comes to naming events and properties, one naming convention should be followed through the whole event-tracking plan. We prefer the object-action framework in combination with snake case convention, but there are lots of options out there.

There is also the camel case, Hungarian, or kebab case. You can also use your own naming convention, as long as you are consistent.

Being consistent and following conventions are two out of the three important Cs when it comes to building event-tracking plans.

The last but not least important is clarity because every defined event name should be clear to anyone who reads the event-tracking plan or analyses the data.

The event-tracking plan needs to be understood not only by the data analyst who is creating it and the developer who is implementing it, but also by the business stakeholders who want insights from collected data to make data-driven decisions.

Documenting each event and property is very important. This way, you know at each moment what data is being tracked. There should be a central document, excel table or a file on GitHub, containing all details of what is being tracked in one place.

To make the event-tracking plan easier to read for business stakeholders, we’ve come up with an idea of making the tracking plan more visual through user journeys and mobile app screens.

We create user data journeys in a tool called Overflow, which our design team sometimes uses for visualizing mobile app flows.

Whenever we make changes in mobile apps, we update our tracking plans and the visualizations of user journeys. This way, it’s easy to see which information we are collecting for each part of the app, as well as to see whether we are missing something or if perhaps the data we are collecting isn’t aligned on all platforms.

This can be crucial because, if something is different, the wrong data is being collected and the results of any analysis we undertake are incorrect.

Everyone wants to know how their app is performing, what are the most and least used features in the app, and why. But to have all answers and to start making data-driven decisions, everybody needs to understand the importance of defining and implementing analytics correctly.

No events set up means no data, and no data means no answers. Consequently, it means you are making decisions in the dark. Even worse, poorly defined tracking can lead to wrong insights, and ultimately, to wrong conclusions and decisions.

Finally, here are the most important takeaways: