Mobile App Launch

↳ TLDR

Background

My entire career at Tempest, I was waiting for us to do a mobile app. It seemed inevitable- since every user I talked to was basically begging for one, and Tempest seemed like a great candidate for a mobile app. In Q4 of 2021, it was finally time to focus on mobile.

Why a mobile app?

Be more competitive in the health tech space- many of our competitors have apps.
There's a big appetite for mobile. In a large cancellation survey, 5% of users specifically list no mobile app as their reason for churn. Anecdotally, every user who has done a user testing interview requests a phone app. In unrelated surveys, most users write in that they would like Tempest to have a mobile app in the section that asks us if there's anything else they'd like to share with us. Roughly 80% of our userbase only uses the mobile app currently.
Increase engagement with push notifications. About 70% of members read daily affirmations every day- in their email. Being able to send them to the app to read these every day will increase engagement more than daily emails.

Risks

A mobile app is costly. Will it have the desired results and be worth the investment? What should the scope be? Should the mobile app have different IA? We have a very outdated backend. How are we going to work with that?

↳ KEY RESULTS

Within the first week of launch:

880 downloads on iOS, 108 on Android
Increased engagement, on par with January engagement (our best month, because of new years resolutions and Dry January)
Weekly overall engagement increased from ~20% to ~30%
Weekly community engagement increased from ~10% to 20%
Positive member feedback
7 5-star ratings on iOS

Data collection update results:

Increased completion rate from 70% to 80%
Avg. time to complete down to 4 min from 18 minutes
Monthly survey completion rate up from 80% to 90%

User Goals

Be able to use a Tempest app, so it's easier to access than a webpage.

Recieve push notifications for my community interactions, and for my daily affirmation.

Be able to use a Tempest app, so it's easier to access than a webpage.

Recieve push notifications for my community interactions, and for my daily affirmation.

thought-catalog-Nv-vx3kUR2A-unsplash-scaled

Business Goals

Increase conversion to 2%, from .9%

Explore sober curious market

Increase retention (our churn and acquisition keep things flat)

Increase conversion to 2%, from .9%

Explore sober curious market

Increase retention (our churn and acquisition keep things flat)

↳ BETA TEST

Test App Using a Whitelabel Solution (Mighty Networks)

Our investors wanted to see a proof of concept- fast, before approving the budget for us to hire mobile engineers for the app. I put together a few proposals for how we could make this work, and we ended up going with the following option: Release a proof of concept free app using a whitelabel solution. The app will target the sober curious market, and people not committed to quitting alcohol. Then, when we get the thumbs up from investors, we create our membership app with an internal eng team.

I put together an app strategy for the sober curious market, including 2 months worth of community prompts and polls, and 3 intro courses.

Data

The app was wildly successful, and helped us quickly get approval to move forward with a member app. It was launched for 2 months, and within those 2 months:

We got 1800+ members (19 referred)
Conversion rate from the app fluctuated between 1.2% and 2%, higher than our Dry January conversion rates of 1.3-1.5%, and much higher than our normal yearlong rate of .9%
Combined community and content engagement rate was 60%, much higher than the ~20% we see on the web app
Got to see what types of posts and content were most popular (polls and simple prompts)
Got to see how much push notifications for polls and prompts effected engagement, and which days and times were most popular for users to log on
Confirmed that users seem less interested in content- only 10% of users checked out any content

↳ USABILITY TESTING

Methodology

To test our assumptions about designs, we made 2 prototypes and tested both with 12 users. One prototype used our existing navigation and IA- the second used the updated IA we thought would work better. We also tested some future ideas.

Results

Hard to pin down what users want in terms of content
Users like suggestions/guidance on what to do
Love affirmation in the app
New IA works fine, no difference in completion of tasks
Overall- good to move forward with new IA on app.
This test proves that we need to put a lot more thought into program and content strategy if we want to see higher engagement and completion rates.

↳ FINAL DESIGNS

↳ INFORMATION ARCHITECTURE

↳ DATA COLLECTION

During this project, I was also working on activation events and engagement data. I noticed that ~40% of users don't fill out our self-assessment survey (which we use for efficacy data), and those users are the most likely to churn without logging on more than 5 times. However, I didn't know why users weren't filling it out- were they seeing that the self-assessment was 80 questions, and then getting scared off- i.e., is the survey the cause of disengagement? Or, were they already disengaged and disenclined to fill it out? Since we were able to get v1 designs done fairly quickly, I decided to spend some time overhauling our data collection.

I found out that a majority of questions we were asking were both unecessary, and not leading to any insights with our users. I managed to cut the self-assessment down to 17 questions, I cut the monthly down from 30 to 13, and made similar changes in other surveys across our platform.

↳ WIREFRAMES

open to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥopen to work ꕥ

Contact

Phone: +1 929 293 7758
Email: sarahhayleyarmstrong@gmail.com