Ask HN: What methods and tools can I use to decrease my app churn?

1) Create the conversion funnel. EG: clicked ad -> installed app -> registered -> engaged with new user flow -> used product part a -> purchased product part b. What parts should they see and use all broken down. As they progress in the funnel they are more engaged.

2) Fill in all the stats of the conversion funnels. This will give you a baseline to answer: Did my modifications help or make it worse. EG: clicked ad -> installed app will be a %50 drop. installed app -> registered was a 20% drop. Amplitude works fine here.

3) Get a steady stream of users in. Ads, forums, social anything. Just a handful so you can see if you are changing things in 2).

4) Setup user testing. (UXCam <- makes VIDEOS of mobile sessions) This will let you see videos of EVERYTHING a person did. Watch each user's behavior 10 times. Over and over. Get to the point where you know what happens next.

5) This is the hard part. Make some change. See if it changes 2). Changes made in the early part of the funnel will change how many people TRY your app. Changes made in the later portion will change how many people STAY in your app.

6) Do 5) until you give up. Over and over. Then if you can't make 2) better quit.

7) If your cost of user acquisition is less than your profit. Run to a VC as fast as possible and pour gas all over it [or not].

5) seems scary to me. Isn't that something you would need the user to opt in to? Or does EDPR not apply to apps?

GDPR does not prohibit tracking. GDPR prohibits not disclosing that you're doing it (and naming 3rd parties that you use).

I like to look at it as a rethink moment for when you build in the mechanism to track. Do I really need this? Is there data I don't need to collect here? Do I need a user's consent?

This is actually a very insightful comment. I don't have experience with the VC part at the end, but tight iteration and measurement loops are key, especially early on.

No idea what your app is, but do you know if the uninstall is happening right after the first open?

I’m asking because I usually uninstall apps either:

- Months later after I notice I’m not using it

- Minutes after opening it for the first time

If you just launched your app, I imagine your uninstall is happening right away.

As an app user, let me tell you what usually makes me uninstall right away.

- App permissions that I don’t want to give. If the app keeps working 100% after I decline, I may keep it. Otherwise, uninstall.

- Prompt for account creation. If the app provides something I REALLY want, I might create an account. If I was just curios, uninstall. If the account creation is with only social accounts (SSO with Facebook or Twitter) and no option for email, uninstall. Others may vary.

- I was looking for an app that does what your app does, and while waiting for a large download, found another app that looks to do that thing better. Either screenshots, reviews, description, something convinced me that it’s better. Downloaded the other one, tried it, decided to keep it and uninstall yours.

- It’s a free app but any useful feature is behind a paid subscription or behind an in-app purchase.

If I had to guess, he might be gating the app with a login. Yes some people might be willing to sign up for a new service, but for me if the app doesn't allow me to check it out before signing up, I almost always uninstall

^ This. You are most likely asking too much of your user and providing little value in return.

The next thing is the user installs it, but finds out how shallow and useless the content / app is for their own use case. You need something that keeps users coming back, what is it?

I think you are fundamentally asking the wrong question.

Instead of trying to understand why 80% of the people are leaving, try to find out why 20% are staying.

Instead of contacting the people who are uninstalling the app, contact the people who install it and stay with it. Ask them why they are buying and what they are using it for.

The people who are uninstalling are not your customers. The people who install and stay are your customers. Find out why they like the app and then use these insights to attract more customers.

Well - in my opinion - having done this day in and day out for last 4 years - the key thing is to look at both users who are leaving and who are staying. We understand the difference between the two and then try to tilt the uninstalling users towards the installing users behaviour. However, sometimes the uninstalling behaviour is specific to something that cohort only and hence both cohorts need to be looked at.

>Instead of trying to understand why 80% of the people are leaving, try to find out why 20% are staying.

That's a survivorship bias. You have to analyze both groups in order to improve the ratio.

Their comment wasn't directed towards improving the ratio. They were suggesting that the focus should instead be on the current users. I imagine this is a result of everything being so new, they think it makes sense to take some time and figure out why people are actually using the app before trying to tackle more directed problems like the uninstall ratio.

In that context, it's really not a survivorship bias. If you want to know about the survivors, it's not wrong to consider the surviors only.

But the OP's question is about improving the ratio.

Right... and then they got a response that they might want to ask a different question along with a suggestion of how to answer it.

Now you are trying to treat that solution that is explicitly meant to answer the new question as if it were an answer for the original question, going so far as to apply logic issues from the old onto the new.

Discussions, particularly ones of this nature, change and shift direction. It doesn't mean that later topics inherit logical traits or something from earlier ones.

Agreed. If a minority of people like the app and you want to appeal to the majority, don't ask the minority what they want because those might be reasons the majority don't care about/dislike.

I'd select 10 random users who abandoned the app and offered them $10 in Amazon ecards for 10 minutes on the phone with me. Then, I'd ask them what they had hoped for when they signed up, and what went wrong.

I think it would help to provide a questionaire form instead. People really really really don't like talking to strangers on the phone.

Speak for yourself. I would prefer a human interaction over a list of questions. Also you'd have more of a chance to get elaboration on things you find interesting.

It depends on the audience the app is targeted at. If it's a geeky thing, phone calls might not work.

Exactly this. 80% churn means you don’t have an app (yet.) or, you have an app, but you are not positioning it correctly. You have the wrong customers basically.

So a few things jumped out at me:

1) It’s going to be a grind and the fact that you increased from 200 actives to 500 actives is a good sign!

2) 10-20% trial signups conversion isn’t terrible until you’ve found the right type of customer you’re targeting. Bigger companies might be getting close to 40% but that is for a product that already has market fit.

3) second other suggestion...email and get in front of your users. Or when users file support tickets (make it super easy for them), use that as opp to get educated.

4) don’t spend money on ads to increase traffic until you are getting more of the type of customer you want and consistently converting them

5) don’t spend more time insrutmenting analytics. You already know what’s up. Just talk to users more and iterate based on their feedback

A lot of great advice here, mixed with some sales pitches and some questionable advice.

App reviews seem to have gone unmentioned - are you getting any negative reviews?

If you are, see if they contain anything actionable or even a hint of what the problem might be.

If not, the uninstalls may or may not be a problem - for a paid app that’s free to install, you could be getting a lot of window shopping. Once people realize there is no free lunch, they move on. If this is the version of events playing out, then these were never your target market to begin with - they were not willing to pay and could have been just trying to get something for free.

Only way to know for sure is, like many other comments stated, talk to them - forums, support sites/emails, even reddit could be a good source. Try to figure out who your customers are and look for the place where they communicate.

I'm in a similar boat with my app and we've found success by implementing the MomTest (http://momtestbook.com/). The premise is that your users are well-intentioned and will tell you what you want to hear, just like your mom would. To work around this, you employ the MomTest- basically you avoid all questions with words like "would you", "do you", etc because they lead to red herrings and reassuring dead ends. You ask questions about their real experiences relating to the problem you're trying to solve so that you can deduce the rest. Your mileage may vary but since you have some existing users, you might consider doing this to extract some useful data about them.

As a general rule, consider number of installations akin to window shopping or browsing in a shop. Most of them will be followed by an uninstallation. The same install/uninstall churn will happen to several similar applications at the same time: when I'm looking for a good app, I might install 10 and uninstall 10, not having found a good one. I just took your product in my hand, wondered about it, and put it back on the shelf. Don't worry about it. I might not be your target segment anyway.

But should you convince me that your app might be worth its while we bump into the next hindrance.

As a general rule, things that make me abort the installation or uninstall right away:

- app crashes or has particularly custom or sluggish UI which stands out from the more modest applications on my phone

- app wants me to sign in to something

- app absolutely requires permissions I don't want to give

These can be born if the application is truly good (useful) and unique (can't go with another app). But most of these points prevent me from figuring out if the app is worth it.

And these all come into consideration only after I have positive expectations of why should I even try your app.

Considering the 20%, it could actually be quite high (depending on the context, app, and target audience of course).

> I have two keywords which I am targeting and have a very nice 50% conversion rate but mostly all of my installs come from keywords that are not listed and get placed in the "other" category.

Maybe your app simply did not do what the user wanted, as they found it with an irrelevant keyword? So basically they are just adding noise to your data as they aren't your target market.

> Currently my app has a couple of issues with a few crashes a day which I am working to resolve but I don't think that is the culprit.

How sure are you about this? Speaking personally, if I installed an app that was otherwise good but crashed that often, I'd likely uninstall it.

Well it obviously has a negative effect but the reports are giving me 10 crashes per day and there are 100 more uninstalls so even if 1 crash = 1 uninstall, why are the other 90 ppl uninstalling...

Ah, I see; I was interpreting "a few crashes per day" as "a user will see a few crashes per day", which I understand now not to be the case. That being said, the sibling comment here suggesting that some crashes might not be reported is an interesting thing to think about, although I don't know nearly enough about mobile device to know how to tell if that's happening.

Or maybe 100 crashed, but only 10 made to report it.

Doesn't this statement:

>I have a high churn rate (install/uninstall ratio) of about 70-80%

Contradict this statement:

>I know its an app with a viable market, and has potential.

Other suggestions:

>I am getting 3-5 free 3 day trial signups

You are giving it away for free, charge people up front, then you will know if it really solves a problem they are willing to pay for. That is when you will know you have product / market fit.

>I use google analytics to track events but they are more of a way to report how much activity my app gets rather than reports on how people feel about my app.

Use Mixpanel or something similar (idk what the Mixpanel equivalent for apps is) to get a more detailed view of what people are doing.

Well I think I have a retention/engagement problem that I don't know how to troubleshoot. Most of the users who sign up for the free trial stick around and pay. So it's not like I am not seeing any sign ups.

Why not make installation part of the free trial?

I really like Amplitude

Mixpanel works great on apps as well..

Consider using some tools like google analyticsz they actually have bindings for apps like .Net

https://github.com/maartenba/GoogleAnalyticsTracker

You can also detect the uninstall event and try to launch a browser to a quick optional survey

https://docs.microsoft.com/en-us/dotnet/api/system.configura...

I’d say try to treat the desktop app as much like a We app as possible from a metrics perspective. Unlike web people have a general disdain for tracking so with all of the info above I’d consider making it opt-in only.

Finally as mentioned above. Post on LinkedIn for some honest reviews of the app, offer some money for their time. At low volume this is easy to manage manually. At scale consider something like https://www.giftbit.com/

Good luck!

Survey on uninstall = where theoretical ideal meets the real world.

Edit: By which I mean, in that moment you do not want to get in the user's way.

Have you tried Fabric.io? It’s a good tool specially for mobile apps. You should also create a Facebook page of your app and make it accessible from app, people like Facebook to communicate via calls or direct emails for feedback.

Why do you think your churn rate is high? Do you know ow what is the average churn rate of other apps or similar apps? (No, your churn rate is not that bad, it is normal in apps). Another good question is the life time value of each customer and how long your users stay in your app (like since they installed it until they uninstalled it). Create an engagement metric for user use case and try to I prove it, more engage means less churn.

Focus on your target market and measure the churn rate in your target market. Forget the rest of the users, those are noise. Develop what you need for your target market.

> without any marketing

Well thats not true, if people (the market) are finding (though a marketing channel) your app (the product) then you have done marketing (posting your app on the store) you just don't realise it.

If you don't plan out your marketing you are still marketing you are just usually doing it poorly by marketing to everyone and no one, you should expect to get a lot of confused non customers.

Summarising what other people have said & adding my 2 cents:

a.) Adding MixPanel, Amplitude or other Analytics is must.

b.) Understand your customer - Are you getting 150 organic downloads? I believe - you are running an ad campaign. Try to make few changes to make your app do organic downloads.

c.) Hire a good UX Designer - Run your analytics & fix the Leakage.

There are agencies who provide consulting on App Growth & Churn - Speak to them. phiture.com, prolificinteractive.com are my favorite ones.

I also consult apps for organic growth & downloads.

Use sentry.io to really understand where and when your crashes and other exceptions are happening.

As others mentioned, offer a giftcard for a brief phone call or chat interview.

You need to check the onboarding funnel as someone has pointed out. But in addition, you also need to look at App "Vitals" like ANR, Crashes, free space, performance

We do this for a living daily. Give me a shout out at tarun@appice.io

Do you demand a registration? Or users can try before giving your their email address?

Do you have a mini tutorial (3~5 'swipes') for the first time users?

Did you try sending a personalized email to users that created an account and then stopped using?

> Did you try sending a personalized email to users that created an account and then stopped using?

If I stop using an app the last thing I want is to receive spam (and yeah, to my eyes this would be spam).

Don’t do this.

yes I have a registration page prior to the main content page. No tutorial animation but I have a button inside the menu. and the emails were not really personalized since I don't have much other than their emails

Maybe try sample content that leads them into registration? You can also try to get them to register after they’ve tried to engage with the content. Users are moser likely to register after they’ve engaged.

This is a good suggestion thanks

Surprised nobody seems to have mentioned it yet, but what about expiring credit cards? Do you send those users an email about updating their cards?

Are you required to put in a credit card for the free trial?

If you can detect a first time install you may want to put a little sales/tutorial in.

I highly recommend checking out useronboard.com to see lots of examples of good and bad onboarding experiences.

https://lesschurn.io/

Something’s wrong with this site. I used Firefox Focus and tapped on the Pricing link at the bottom, but it kept going back to the homepage. I finally typed “/pricing” after the domain in the URL to get to the pricing page.

How are you getting the installs in the first place? Were they accidental?

Is the performance of the app bad?

support is a huge factor in reducing churn. can they reach out and is your team/you responsive? if they complain about crashes can they get a credit for the month while you fix, etc? good support will cut churn

https://meetings.hubspot.com/brett38

Sign up for a meeting, happy to chat about retention/growth/whatever! Let me know if you need a different +/- time zone and will be happy to accommodate you.

May I ask what your app is?

During uninstall take them to a page where you very simply and sincerely ask them for feedback. Promise that you will read it and respond to it. Don’t require an email on that form though so they can answer anonymously if they want to.

Great suggestion but from what I've read there really isn't a way to detect uninstall events on Android.

Can you infer uninstalls from heuristics? If you're collecting activity metrics for app interactions and you get a complete dropoff for a given user, could you assume they have uninstalled? Even if they havent, checking in on decreased engagement might be an opportunity to get them before they uninstall.

Not sure what the timeframe is between a user trying it out and subsequently uninstalling, but if there is a bit of a time gap, the decreased interaction analytics could be a used as a leading indicator of pending uninstallation and used to trigger a win-back attempt?

So much free advice here. Pay a UX consultant to help you.

Is your app, appsee.com?

In addition to GA and/or Mixpanel, I recommend fullstory.com for web based apps to see individual user sessions in very high fidelity.

Don’t be creepy. Users don’t expect you to be able to watch how they interact with your app (and what data they enter in it) so don’t.