But what I'm really curious about is how you actually put the survey in front of your users: Do you pop it up in-app, or link it via in-app banner, or send it via email/SMS/push?
What works to get lots of responses? What works to get high-quality responses?
That way, the request for feedback itself is not a negative, jarring experience. This is what I did on https://www.joyapp.com and had very positive results.
- Respect the user and not send any annoying requests for suggestions -> Getting feedback should be a very minor point of your app, assume the user does NOT want to see any requests for information
-Think like the user, use the app extensively and try and make it as perfect without their feedback
- Offer a discrete email for them to contact you (usually within the help page)
- Add the option for reviews & feedback (via the Apple API)
I get quite a bit of comments via reviews and also the occasional email
Because users have to actively reach out to you, usually the quality of the feedback is very high
Currently rated 4.8/5 on the App Store, so it’s working for me. That said its not a complex app, so I haven’t had as much need for user feedback
NPS and similar are old-school corporate CYA tactics. Today you can look at the attributions for referrals and work with what people do not just say.
For improvement (e.g. you want to increase on-boarding completion rate) you need behavioural interviews.
Here are my latest channels: hubspot, email, intercom, inapp. Why I chose one or the other was determined by target audience (who I want to gather info from) + relevance (how close to the moment I want to find out about can I get) + tech capabilities (what is possible technically speaking).
Prior to this decision there was extensive behavioural mapping and profiling. It is the only way you can get quality responses, as user feedback is mostly qualitative.
Let me know if I can help
We also send (often a different) survey email (along with product updates) to folks who've subscribed in case we expect to learn something else than from the responses in the telegram group. In our experience, with our audience, emails don't work as well as Telegram.
What we also do to gauge user-interest for an Android (userspace) DoH client and firewall we build is:
1a. Include a link to Telegram community in the about page prominently.
1b. After every update, shows a changelog to the user with a button "suggest features" that emails feedback.
2. Contact email features on all our webpages.
3. Google the past 24 hours for keywords related to our app to see who's discussing it, and if it is a forum like reddit, leave a comment and personal email (as opposed to generic support at domain dot com) encouraging the user to contact.
4. Newsletter to collect potential customer email-ids. As mentioned above, survey are always sent out to them once in a blue-moon and only with accompanying significant product updates.
5. All incoming to generic mailbox are replied-to from a personal email. Emails always end with a remark asking the user to not shy away from sending bug reports, queries, and suggestions.
6. Monitor related subreddits where similar products/apps are discussed, and post replies only after the thread is off the front-page to avoid derailing a hot discussion with self-promotion (which gets you banned on most subreddits).
Our support form lets you do feature requests, information requests, as well as support request. We review and answer every one.
People who submit multiple and will engage with us we put on a list. Many of those people make it into TestFlight and we have a tight feedback loop with them. They can opt out any time.
But we also have a lot of users with a TON of time in the app (some of our people are in the app 6+ hours a day). So they hear about stuff pretty quickly and we try to respond as best we can to fix and improve things.
But really, just responding to feedback requests and figuring out who has good ideas and can work with you on those... that's the biggest part aside from the time it takes to manage those relationships.
I plan on keeping it this way for quite some time, until it becomes absolutely impossible to scale further.
I usually just ask "How are you using it?", or "Is there anything that didn't work like you expected it to?"
Or I go "And if you want to do this, you click here" and there's an error message that we used to click on to dismiss unconsciously when using the product.
My point is, regularly doing videos that demonstrate how awesome the product is reveal how much the product frigging sucks, because there's cognitive dissonance. You find yourself saying "Oh, you can easily do this" but the application as it is clearly shows it's not as easy or reliable. I do this for new branches and features because it puts me as a user². The videos reveal a lot of crap that became invisible to you.
One other thing is analytics. I sent a message to a colleague of mine asking him questions about why he was doing certain things with tokens.
Track non consumption. Non-consumption. Non-consumption. Are the users able to accomplish core tasks? No? Why? Do they have trouble finding the feature exists? Is it burried? Can the users find the feature, is it easy to use, can they use it consistently, is it reliable?
Dogfooding. Do you use the application or is it for someone else? For example, I asked a ride-sharing app team about some use cases: the ability to block drivers I never want to ride with again (one of these drivers got into three road accidents in a single ride, went against traffic, violated roundabouts, and texted while driving. I don't ever want to ride with them, and a review/rating is not enough for me. I have taken multiple rides with that person giving them the benefit of the doubt, but they frequently get into accidents other than that instance. I want to make sure I never see them again, and I figure there are drivers who want to block certain riders as well because a driver claimed they were far until they saw me. They said my phone number looked similar to someone they vowed not to take again until they saw I wasn't that person). This use case is obvious if you ever use the application.
Otherwise, chat seems great for early stage (< 500 DAU). I didn't even use a real chat, it was just a giant, non-live wall they could post on. You get real, low friction feedback. I tried moving it to discord, but joining a discord server with one person is too intimidating.
Past about 500 DAU, the text moves too fast though and you start getting trolls.
If you ask everyone, everyone will have a different opinion and you'll drown in the noise.
I have a contact email that works, and certain 'power users' never stop bugging me but their feedback is invaluable. Also helpful is monitoring public comments etc. A lot of times that brings attention to issues that no-one has thought of reporting.