A well-known Swedish media publisher IMP needed a digital transformation to satisfy the needs of today’s digital world. Focusing on Millennials as a target group, we were challenged to create something extraordinary to defer in the already saturated market. Through many hours of research, market analysis, and surveys – we found the sweet spot that will capture Millennial’s attention and satisfy their cravings with meals based on their diet preferences, allergies, and needs.
PixFood, the AI-driven food assistant, was born
PixFood is an AI-driven digital food assistant that uses advanced image recognition and machine learning to provide the best recipe choices based on the everyday routines and habits of millennials.
The app provides a simple & fun user experience that starts with the user taking a photo of any ingredient they want to cook and getting tailored recipe suggestions that also take into consideration the time of day, the season, and their dietary preferences.
All of the recipes in the app were carefully selected and they’re featured with high-quality step-by-step videos created by professional chefs and produced in-house.
PixFood the app that gave a new approach like no other food app on the market was livings its dream. The word quickly spread, grabbing international recognition by the world’s most prestigious high tech publishers such as TechCrunch, Entrepreneur, Lifehacker, and more. The things started moving in the right direction resulting in Nr.3 product of the day at Product Hunt, which helped us lower the cost per acquisition to 0.5$, racking up 30,000 downloads!
And then the challenges appeared…
And even though the acquisition was going strong, PixFood was having challenges with retaining its new users.
To understand why this was happening, and pinpoint the underlying issues that we need to address – we built a tracking plan that outlines key user actions and properties and then used Mixpanel’s Product Analytics to get a deeper understanding of the users.
The power of Product Analytics
Things started escalating quickly and in this dynamic world, you need to have the best tool on the market to understand the data from your customers. Luckily, we have implemented Mixpanel to track our customer actions and behaviors. The power of product analytics gave us an advantage in understanding what’s happening in the app and where and why the newly acquired user’s drop-off.
Defining clear action points that need to be done:
- Analyzing current flow followed by a Marketing funnel.
- Finding relevant data for current users.
- Defining the pain points and main drop-offs.
- Finding the exact pain point in each step of the funnel.
- Creating suggestions and drafts created by indicators from the drop-offs/pain points.
By creating the possible scenarios with Mixpanel’s Insights, Funnels, Flows, we had a clear picture of what’s happening in the app. This provided us with the data and with the right insights to start making data-driven decisions.
Creating different segments for the most frequent flows.
- Segment 1: First App Open users who got the right recipe suggestion.
- Segment 2: First App Open users who followed the current flow using the ingredient recognition screen, type ingredient and predefined ingredients.
- Segment 3: First App Open users who opened the app and made NO ACTION
- Segment 4: Returning Users who:
– Got the right recipe suggestion
– Followed the current flow
– Made No Action
With these segments, we covered most of the scenarios and we could start acting on them. We defined 4 critical moments that needed to be improved.
Critical Moment 1:
In general, users don’t grant access to the camera if they skip the on-boarding flow – which they do (almost 2/3 of new users).
What can we do about it?
=> The users must be informed about why we ask for the camera.
=> We must find a better way to provide that information (not by using a system alert only).
*** Emphasize: Camera allowance must not limit other functionalities.
Critical Moment 2:
First-time users explore in general, so it is expected to have a big drop-off on a screen where we demand some action (snap, type, etc.). But, the drop-off there doesn’t decrease a lot when the users return. This is opposite to what happens when the users go to Recommendations first – the drop-off is somewhere between 35-40% for both returning and first-time users, which is not satisfying, but still far better than the other one.
What can we do about it?
=> Direct the users to where they make smaller drop-offs and increase the engagement there.
=> Then simultaneously provide value from other functionalities to let the user give permissions, use search, provide preferences, etc. which will eliminate drop-offs that may happen because of unexpected demands, lack of information, misunderstanding value, overall miscommunication, etc.
Critical Moment 3:
Users don’t register – neither first-time nor returning. Not even gated content itself influences the registration.
What can we do about it?
=> Find a way to let the users register (limit the content you provide) which doesn’t demand a huge effort to make them get what they want (to pay, let’s say).
=> Then use in-app messages and other methods to increase the chances for engaging users within the app.
Critical Moment 4:
Users don’t cook even when they open recipes. This may be caused by inadequate content or by frequent app crashes.
What can we do about it?
=> Find a way to gain information about what users don’t like about the recipes (by using in-app, marking recipes, surveys, etc.)
*** Eliminate the scenario where the user searched for a recipe, got one, got back accidentally, and then is made to search for it again.
The outcome
The insights from Mixpanel’s analysis provided us with the means to build the perfect onboarding flow, targeting users with personalized messaging across all channels, at the right moment of their user journey.
The outcome of the process was amazing!
- First, we came up with a completely different action-based onboarding flow that helped increase the sign-ups by showing triggered tutorial messages at the right time in the user’s journey.
- Second, we increased the overall conversions, by sending personalized messages to the users at the right moment, at the right time, to encourage the desired action which resulted in increased retention rates.
- Third, by sending different messages we found out which content works best, which helped us increase the customer lifetime value and the engagement rate. Using the retention feature by Mixpanel and the segments we have previously created we found that different groups come back in the app for different reasons. Having that in mind we were serving the most engaging content to each group which resulted in decreasing the churn rate.
Having Product Analytics helps you understand your data. Having the right data helps you understand your users. Having the right expertise helps you use the insights to make data-driven decisions for the future of your products & your business.
Acting on your product with data-driven decisions will help you increase your LTV and optimize your product for your customers’ needs which will result in increased conversions, retention rates, and decreased churn rate, which means a successful and profitable product.
It sure did for us:
- x10 increase in Monthly Active Users
- 1,000,000+ personalised messages sent
- 15% increase in sign-ups
- 30% increase in retention rates
- 12% decrease in churn rate
Want to see Mixpanel in action?
At Alite International, we specialize in using data and technology to deliver human-centric experiences across the digital landscape.
If you want to know how we can help you achieve the results you want with behavioural data, feel free to reach out and drop me a line at filip.nedelkovski@alite-international.com – I’ll be happy to help! 🙂
