How to Get a 4.9-star App Store Rating
If you have an app in Apple’s App Store, you know how important the app rating is. Consumers will compare it to competitors; search ranking inside the App Store is impacted by it; and the cost of your app install ads moves is quite dependent on it.
Securing and maintaining a great app store rating is very hard; it takes systematic effort over many months. This is exacerbated by the fact that some factors that consumers are rating you on are outside of your reach as a product manager or a marketer.
For example, if you’re Doordash (and more broadly, if your business sells physical goods), consumers that are upset with their cold dinner will leave you 1-star reviews, even if the root cause behind the late delivery is a traffic accident - something far beyond your control. Given that, by my estimates, a good 10% of Doordash’s orders are late… it’s really quite a feat that Doordash boasts a 4.8 rating.
For Hungryroot, a personalized grocery service where I work, the app is an important part of the user experience. And we also sell physical goods that can get damaged en route. Our Operations teams do excellent work, and yet, sometimes issues happen. So the app team has to work extra hard for the 4.9 star rating that we’ve earned over the past 2 years. Here are the core tenets of our playbook.
1. Eliminate app crashes and severe errors
A crash while using the app is perhaps one of the worst things that can happen in your digital experience. Typically, it results in data loss (“I was just about to place an order, now I have to start over again, argh!”) or at the very least, user confusion and disdain (“what just happened?”). Avoid crashes like the plague.
ACTIONS:
Set 99.9% crash-free sessions as your target. Use a tool like Crashlytics/Firebase to monitor crashes across app releases.
Set up a continuous integration, with builds and top user flow tests automatically kicking off after a developer merges new code. The cost of a significant issue with an app is much higher than with a website, as you cannot patch it instantaneously. Thus, automation infrastructure for your app is even more valuable than for your website.
Set up severe app error logging in a tool like Datadog. Configure alerts in case of a spike. Systematically reduce the error count through regularly scheduled sprint work.
Any time a spike shows up, treat it like a potential sprint interruption. That is, have a triage conversation with the relevant stakeholders (engineering lead, product person, engineer responsible for the crashing feature, possibly also customer service lead) to decide whether the issue is bad enough to warrant a hotfix.
2. Engage with reviewers
Anyone who’s taken the time to write a review in the App Store has had a pretty emotionally moving experience (typically, a negative one).
ACTIONS:
Set up a process where customer service reps reply to every single AppStore reviewer with the mission to (1) make them feel heard (2) mitigate issues whenever possible (3) understand if there’s a yet-unknown bug in your app. You’ll be surprised how often you can turn a detractor into a fan by engaging quickly and being helpful. A tool like Sprout Social supporting that is well worth the money.
Create a “social listening” process: a monthly summary of app (and non-app) feedback that your CX and marketing teams collaborate on. This should be one of the inputs to your product roadmap.
Create an easy route for negative feedback. Customers will only head to the App Store reviews section if there’s no other recourse - they know that in many cases, their complaint there won’t get answered, and they do it mostly to blow off steam. If you create a route for them to complain from within your app, in the context of the action they’re trying to perform, they’re far less likely to resort to expressing their dismay publicly.
What’s a good moment to solicit negative feedback?
Doordash: immediately after every delivery. Ex: “How was your delivery?” Thumbs up/down; if thumbs down, send straight to customer service flow.
Expedia: immediately after check-in at your hotel. Ex: “How was your check-in?” Rate 1-5 stars; if less than 3 stars, go to customer service flow.
Delta Airlines: along with the “flight delay” notification. Ex: “want to connect with a representative to get help with your itinerary?”
3. Prompt consumers that just had a positive experience to rate your app
The founder of modern marketing, Robert Cialdini, teaches us that reciprocity is hard-wired into human behavior. That is, if someone does something good for you, you’re strongly compelled to return the favor. Thus, the moment right after you’ve created value for the customer is a great moment to ask them to do something for you as well - like rating your app five stars.
What are some examples of those memorable moments?
Delta Airlines: after your plane lands early.
Domino’s pizza: after you redeem your rewards points.
Uber: after you rate your latest ride 5 stars.
ACTIONS:
Work with your user-facing teams (market research, customer service, voice of the customer) to identify moments of unbridled joy associated with your product. They surely exist if you’re still in business, look for them :-). Organize product A/B tests where you’re trying each of those moments.
Bonus points for figuring out a strategy around what to ask for during those moments: prompting for a rating is just one thing; you could also ask them to refer a friend, or do another similar “favor.” Experiment to find the optimal route.
Constant Stream of Positive Reviews
Detractors are powerfully motivated; there will always be a steady stream of bad reviews. Your job is to ensure that a stream of positive reviews is also consistently coming to counteract and ideally overpower it, pushing your average app rating up. Thus, a couple rules of thumb to follow:
1. If you stop actively pursuing positive reviews temporarily (for example, if your prompt stops showing up due to a bug), your app store rating will take a dive, and it will be extremely difficult to recover. If you allow the heightened inflow of bad reviews to continue unchecked - for example, because of a crash you’re waiting for 2 weeks to fix - you will be in a similar pickle.
2. Do some forward modeling of ratings, that is, predict where you’ll be in 6 months if the current trajectory holds. For example, you may be at 4.8 average now, with tens of thousands of reviews, but the last few months of ratings have been more negative, and at that trend, you might end up at 4.7 soon. A good moment to react to that is now, not when you’re already at 4.7.