Ask HN: How to effectively get feedback from users?
Getting feedback from customers is crucial but it seems really difficult to get people to answer a survey or get on a call.
I mainly use email to ask for feedback and the response rates are terrible, 2-3%.
Are there any effective ways of getting feedback from customers? Tools or templates?
90 comments
[ 3.0 ms ] story [ 151 ms ] threadIf you don't have any, log user actions when they succeed or fail to do something. Even a simple log can tell you many things.
It also tells you if they're failing to discover or use a certain feature, or if they're using something in an unintended way.
You can set up Slack channels for your product to help users troubleshoot issues. We have workspaces where users can get help. Whenever the 20th user complains about something, what to do next becomes clear. We tend to focus on issues with high frequency and high severity as opposed to say stylesheets, but it's because the field is struggling not for lack of good CSS. But it may be different for you and you may have a product where the value of the product is the better interface. Adjust accordingly.
- [0]: https://posthog.com/
- [1]: https://github.com/PostHog/posthog
At the early stage my experience has been:
1. You should install SmartLook, FullStory, or similar to watch users. Qualitative feedback beats quantitative b/c your raw numbers are too low to be useful.
2. Identify key points in the customer journey when you want to collect feedback. For B2B SaaS I ask for feedback via email at these stages:
- When a customer signs up for a trial
- When a customer converts to a paying sub
- When a customer doesn't convert
- After 1-2 months of paying
- When a customer churns
First I'll make sure that I get system-generated emails when these events happen so I can send them manually.
I'll eventually automate sending these emails once volume gets too high. I'll automate them by sending from our app or from a tool like UserList or Intercom.
And when I get feedback I centralize in the tool I'm building (www.savio.io) and use it to prioritize what to build next.
It's pretty effective at getting very early bug reports, gathering interest on potential features, and their initial use was for feedback gathering. Random users won't give you feedback however until they came across a bug or a need though in my experience!
* If you're pre-product/market fit, just reach out to users via email, in-product messaging etc. You can incentivize them with Amazon gift cards as well. At this stage nobody actually needs your product so you'll have to do it by hook or by crook.
* If you're post-product/market fit then people actually _need_ your business. Once you make it clear that you're going to action on their feedback they're highly incentivized to help you out – so use the same strategies as above but make it clear what you'll use the feedback for.
* Set up automated surveys, especially ones that are unobtrusive in order to not create a shitty user experience – this allows you to baseline your product's effectiveness which is the first step towards continuous iteration. Eg last quarter we were at 3.3/5 satisfaction, let's get to 4.0/5 this quarter. These surveys are also a great way to find customers to interview.
* FullStory and Hotjar both allow you to view user sessions, highly recommend them as well.
The most important point: if you're an enterprise business, I highly recommend identifying the "best" customers that you have, building relationships with them, and favoring their guidance over others'. Once you hit scale all of your customers will want to give you feedback, but only some of them will have the wisdom/intelligence/creativity/whatever to have a great sense for what _you_ should build for _your_ business to succeed. When you find these customers, get them onto a customer advisory board / meet with them a lot as they will help you in the art of pulling a great product out of your market.
If you value your user's privacy (as well you should if you ever want to do business in the EU) do not be tempted by this privacy-hostile path.
I fully disagree. Consenting to a usability study means I want to provide feedback and actively state what I'm doing, why I'm doing it, and what I didn't expect.
It does not mean I want you to see what I do, where I leave my mouse, what I click on, what input I type -- and what input I then delete. If I wanted that then I would take screenshots or a video and share that.
> I fully disagree. Consenting to a usability study means I want to provide feedback and actively state what I'm doing, why I'm doing it, and what I didn't expect.
This is fine for an UX study that conducts in-person interviews as _one_ of the methods. This method is almost useless on its own, mostly because people behave differently when they know they're being observed (eg they do things in a more "optimal way" in order to avoid looking stupid in front of the person conducting the study).
Also, while active feedback is valuable, it can be harder to interpret. People say A, but mean B and then do something between A and B or just do C. Not because they're malicious, but because that was what came to mind when they tried to express something or just didn't know better, so there's an unintentional misalignment between intent, action, and what was expressed.
> It does not mean I want you to see what I do, where I leave my mouse, what I click on, what input I type -- and what input I then delete. If I wanted that then I would take screenshots or a video and share that.
This is exactly the type of thing you (as a designer) _want_ to know in order to identify friction and remove it. Obviously, you track what the user types only under certain conditions. An example would be, a number input: do user always type the same/similar values or increments of some values (isolated per user), should the input offer pre-defined incementors or buttons that just fill-and-submit it instantly (you'd do AB tests to see which work better then), etc.
I fully understand your concern about privacy, but not everything is run by reprobates who sell user's data. Sometimes "observing" is the best/easiest way to learn how to improve a user's experience.
You would prefer not to participate in such a study, and that is perfectly fair.
That doesn't mean its an ineffective method for observing how users interact with software.
I've been surprised many times with (in-person) usability studies where we ask users to perform certain tasks with the tool. It's a great way to improve UX.
Did anyone really need the early AirTable? No, you could just use Google Sheets. Did anyone really need WhatsApp? No, there was texting already. And so on.
In theory everyone should stay far away from building "nice to have" products and only solve hard business problems but that is a super high standard that even most companies that are already out there probably don't meet.
Of course, getting feedback and adjusting your vision is super important though and sometimes a radical pivot is needed.
Yes, AirTable added additional benefits compared to Google Sheets. So did WhatsApp. That's why they succeeded in the first place, their first (public) iteration was useful enough to gain initial traction.
I'm not arguing that everything is useless unless they solve "hard business problems" (which I'm not even sure what that refers to), but if your thing doesn't solve anything at all, it's either entertainment or the wrong thing to build.
Since we know that 3 out of 10 users who receive this offer will accept it we segment users by industry / feature usage etc. and send offers to talk to them to listen their stories and feedback.
https://amzn.to/383bo92
It talks at length about the challenges, especially, of getting _honest_ feedback from prospective users. And it has a lot of ideas to that effect.
https://www.amazon.com/Mom-Test-customers-business-everyone/...
"We are very interested in how you are getting on with <product>. Please email us if you have any suggestions about how we can improve <product>."
But that only works if you are getting sales.
Also the more responsive you are to customers, the more likely they are to report bugs or make suggestions.
I wrote about my experiences getting feedback here: https://successfulsoftware.net/2008/04/28/getting-customer-f... (old article, but still relevant)
My assumption being that if you're sending unsolicited mails so quickly, then I should be expecting a truckload to arrive by the end of a year. In one of those cases I felt vindicated as I received a bunch of huge exit interview type replies.
That would be an incorrect assumption (in my case, anyway).
Firstly though, you need to have an idea (doesn't have to be perfect or 100% right) of who your target segment is. These are the people whose problem you are solving - and therefore initially are going to hold more weight with their feedback.
Once you have identified these users - use a combination of current users and potential users to get feedback from.
Some ideas:
* Email (like you've been trying) - Try and be candid and frank. Put yourself in their shoes - why should they respond? If it's a generic email their probably going to be of the mindset of either 'Someone else will do it' or 'I can't be bothered'. If you make it unique and more personal you're going to have a higher reply rate - but obviously a higher time commitment. I'd also suggest pushing for feedback calls rather then just messages. A call with a target user is worth its weight in gold.
As other users have suggested - offering a giftcard in exchange for a feedback call is also another valuable route if you can afford it.
* LinkedIn: Depending on your business (B2B or B2C) you can follow up with these users and potential new users on LinkedIn. Again this comes back to being somewhat more personal. Explain who you are etc.
* In app messaging: If they are constantly using your app - ask them while they are there in the midst of it. Don't be spammy though (this is approach we take mainly with delighted).
Lastly, once you have gotten your feedback - don't just take it at facevalue. Try and really understand what problem they are trying to solve as the solution might be completely different then the one they were thinking of.
E.g. if you were building an analytics app and someone gave feedback that they want to be able to export the data. If you take that at facevalue you will most likely build a tool that exports their data. However if you dig deeper and try and understand why they want to export the data, you might be able to solve a bigger issue for them AND for more users. E.g. if they come back and say they want to do XYZ with the data beacuse that means they can then gain ABC insight from the data - you might be able to add the ABC insight into the dashboard instead, skipping the need for export and customer work.
If you want more of a hand feel free to reach out! I'd be happy to help you out!
[1]: This is actually a problem we are trying to solve with Delighted so I feel your pain - https://getdelighted.co
For qualitative data such as user interviews/calls the solution for most teams is either to spend a lot of time on find users to talk to and set up calls or to pay companies like usertesting.com for access to their panels.
These tools don't allow you to get quick (and for most teams affordable) access to your own users for user interviews though, which is something that was enough of a pain point within my own team that we've built a tool exactly for that (https://tryribbon.com).
The Mom Test is also a great read on how to get useful user feedback in general (https://amzn.to/383bo92).
We have found that our customers collect up to 35x more feedback than before once they add the widget to their websites! It does depend on where you place the "Give feedback" button though — the more visible it is, the more feedback you'll collect.
Beauty of it is they sneeze, we know. Everyone in the team kinda knows what's important what's not. we can confidently say things like "nobody cares about that fancy feature". It super valuable for a startup to get immedeate feedback. Especially valuable has been knowing - the first things the notice when they start using: where is sort? how do i add team members? do i have to prepay? - feature requests: it's painful to keep closing windows, need the ability to adjust windows, I also want to do this other thing you guys havent thought of. - Bugs: why am i seeing double messages.
Disadvantages are that some customers expect immedeate answers, unrealistic expectations. and its cringey to lay down the rules and set expectations every time. no matter how many times you do it.
Our take is its 80% good and 20% bad. As we grow the company, the idea is to keep doing this for a "representative sample of customers" . Customer feedback is gold. If you can setup some kind of process like this, its great.
* Hitting up active customers on Twitter, just writing a short message: "Hi xyz, just saw that you are using Splitbee quite actively. Do you have any feedback for me, that would help a lot"
For this method I got answers from 9/10 people with amazing feedback. Very detailed and long messages. Be sure to target "power" users.
* I sent automatic emails if someone embeds splitbee successfully to their page & a mail if they don't do it within 24 hours. People often engage with them and explaining why they did not embed it yet. Also pretty helpful
* We added the feedback.fish widget to all of our https://splitbee.io/docs/embed-the-script and in our app. We get about 8-10 submissions a month which is quite a lot for our size.
Secondly, (maybe it's mentioned I've read all comments) you need a community and keep it active and read the on going discussion on how people think of your product
- I have installed a chat popup on my landing page and I always try to answer any question within 1 hour, and then follow up with those users afterward.
- I have a feedback portal (similar to canny). Whenever anyone leaves any kind of feedback, I write them a thank you email and try to elaborate more on the feedback they left, no matter if this is negative, positive feedback or feature request. For example, if they ask for a feature, I tell them when I plan to develop it.
- Whenever user purchases a license for my app, I also write them a thank you email, and ask for a small feedback. I try to do a short investigation on what user is doing and include some personal touch in it. Usually, 50% of such emails get an answer.
- If I see some bug report coming from a known user I immediately send them an email explaining why that bug has happened, how and when I'm gonna fix it.
I also keep a page in my notes called 'Followups' where I put the user's name and the reason why I should follow up with them. For example next week I'm preparing a new release of the app that has changes which I know are used heavily by a couple of users and I will reach out to them this week, give them an alpha build and ask for their feedback. Whenever I get an idea of how I can bring someone's attention to the product, I add their name to this list.
The responses on here are reasonable but a bit short sighted. What works in one context will not work in another.
Heaps of information on how to get started in this blog: https://userresearch.blog.gov.uk
A discord server is useful for far more than most people understand. You can show daily progress; show new features; highlight community contributions; have a centralized bug report channel; and the other users will start solving other users' problems when it attains critical mass.
The nit is "critical mass". But if you make something that people find remotely interesting, people will join. Broadcast what you're doing on twitter, and eventually you'll get noticed.
Surveys and cold calls are like trying to make fire by banging stones together. Tools from a different era. Sure, they work, but I've found the above process to be far more effective.
It helps to make your discord a general topic, by the way, rather than a specific product. Our ML discord recently passed 1,000 users. https://discord.com/invite/x52Xz3y
Community building is tough work, and I've had several failed attempts. You shouldn't expect yours to take off immediately. You should do it because you like it, and eventually something will work.
But it depends entirely on the business. Most businesses don't need a community. But I can't think of a single business with a community that would have been in such a strong position without it.
Our idea is to keep the perceived effort to give feedback as small as possible and that works quite well. Also, it can embedded and triggered in any software to collect feedback right in place.
Feedback is collected publicly by default. Here is our own page: https://www.stomt.com/stomt but of course you can also turn your page private.
Go where your customers already are and become a part of their community. Be a long-term person in a long term business. If they don't have a place, make it.
If you have that, you're on your way to Product-Channel Fit [1].
Low level:
I've also sent pointless emails and made useless phone calls and most people default to doing nothing when you contact them.
So:
If you think phone/email are where your people are, you need to follow up a lot. Steli Efti has lots of great content on this type of thing [2]. This is an unpopular view on HN because most people see this as spam, but this is what CRMs are built for.
[1]: https://brianbalfour.com/essays/product-channel-fit-for-grow...
[2]: https://blog.close.com/follow-up/
When you send an email and ask user to give feedback they don't really have an incentive to give feedback, maybe giving Amazon gift cards etc.. can increase the likely hood of reply.
The placement of give feedback link can be really important.
Maybe keep in near places where you are more likely to have issues or you want feedback from.