Product managers know that if they want to keep their app from landing in limbo, it's important to keep the rating high. In today's digital world, your app is only as good as its ratings/ reviews on the app/play store.
A good rating goes a long way but most importantly helps you in:
Increasing your daily average installs (Both Google/Apple let apps with higher ratings come to the top/ recommended section)
Subsequently reducing your CAC (making way for organic installs)
Big social proof for first-time users to pick you over your competitors (strengthens consideration)
Most common hacks to improve the rating:
There are a few tried and tested ways to improve your app ratings. Over the years, brands have figured out ways to pick 'AHA' moments with their customers to ensure a rating that more visibility on the stores, more downloads and higher revenues.
Facing pandemic, when millions of people turned to music to pass time at home, at Wynk Music we felt like a huge opportunity came our way. We immediately started with the following:
🚀 Cohort based campaigns - Instead of running a mass campaign for all our users, we identified tiny cohorts that were experiencing 'AHA' moments to run these campaigns. For us, these cohorts were:
Users who downloaded the Cohort based campaigns allow you to find high-performance buckets of people who are enjoying your app. It also allows you to target only good experiences as opposed to running a campaign on a full base. These were trigger-based campaigns and only initiated when the event occurred. 10th song.
Users who set their 2nd hello tune in a month.
Users who streamed more than 100 songs in a week
Cohort based campaigns allows you to find high performance buckets of people who are enjoying your app. It also allows your to target only good experience as opposed to running a campaign on full base. These were trigger based campaigns and only initiated when the event occurred.
🚀 Different campaigns for android and iOS - It could be a huge stereotype or a general myth but iOS users are slightly more sophisticated than android users, especially in India. A blessing in disguise, this gives you a lot of freedom to use quirky messages/ songs for Android users while pushes you to keep it classy for iOS.
Below is what we used for Airtel Xstream app.
🚀 Don't send everyone to play/app store - We didn't send everyone to the play store directly. What if they had something to complain? For everyone who gave us a rating of 3 or <3, we took them to a Typeform where we asked what spoiled their experience.
The responses were duly collected and sent to the product team. For those who rated us 4 or 5 were the ones who were sent to the play store to rate us. This small change can help you bring your weighted average to the higher side.
🚀 Remember to pause - Make sure you ask a user to rate you only once in maybe 6 months. Therefore, when you pause/ restart your campaign, remove the old cohort. You can ask your team to create an event for 'app_store_rating_done' for easier identification.
While exploring new ways to further amplify our rating engine, we stumbled upon RFM.
By this time, our app had already ranked #1 on Google play store, we stood at a rating of 4.2 and we were getting significant gains in our daily average installs.
The recency, frequency, monetary framework, also known as a the RFM model is based on the hypothesis that '80% of business comes from 20% of the customers.' Truly so, it helps you identify these cohorts that fall under the category of being a recent user, a frequent user and those with an affinity to pay. You can read more about this model here.
To use this model, you start by giving inputs on the time window, which model you're picking (RM, RF, or RFM) and the event for which you're making this cohort. For us, it was play_start (when users play any content on the app).
If you run the RF model on your marketing automation tool (we used MoEngage), it will show a table like this. To pick an audience from either of the cohort, just hover over it and save it.
From here on, it's pretty straightforward. You know your loyal customers are your biggest bet - they have a high frequency and recency on the app, pick them. Find a creative way to ask them to rate their experience. Platforms like MoEngage first grade, further analyse, and then segment these bases in order to engage customers as distinct groups, ensuring that the bifurcation is accurate to the last mile.
For a great response, users sitting in loyalists + potential loyalists are the strongest of all cohorts and will drive you the maximum results. More than anything, this also gives you an opportunity to exclude users who're not frequent on the page and have higher propensity of giving you an average rating or no rating at all.
After successful completion of using RF for running the rating campaign, we touched 4.4 for Airtel Xstream on play store and surpassed top OTT giants like Netflix and Prime.
What did it do for our product afterwards? Well, we saw a huge spike in our daily install rate from the play store given our ranking improved. Another domino of landing our 'good' users to the playstore led to great reviews that helped us gain the confidence of new users.
As for Wynk Music, we unlocked a rating of 4.6 (highest ever for us). We ranked amongst the top 10 free apps recommended by Google in the music and entertainment category, and just like Xstream, the installs were growing by the day.
Appstore ratings are crucial discovery, downloads, and in-app purchases - ours mostly talked about the user discovery, feature exploration, and all things free and premium that users loved on the platform. If you don't feel confidant enough to use RF model for rating or simply do not have that kind of base at the moment, stick to the hygiene practices we discussed in the beginning and you'll be just fine.