“Let’s see what the highest voted features were on our feature voting site.” – Said no great product team, ever.
Whether a company is using the new hotness of a Canny, or the old clunky of a UserVoice site, public feature voting systems are an example of terrible product management practices.
Feature Voting is not talking to customers
I’m a HUGE fan of founders and product managers talking to customers. In fact, I’ve written a variety of posts to specifically help more people confidently and effectively “get outside the building” to learn from their customers to build better products.
Unfortunately, feature voting apps are the kind of shortcut that make some people think they’re getting customer input, when really they’re making a mess.
I’ve had a personal disdain for these tools for many years, as I’ve tweeted about issues a few times, and had many conversations with other founders and PMs about it. And today, I’m finally pulling together a comprehensive case for:
- Why great product teams would never dream of using a Feature Voting tool
- All the reasons feature voting leads to a worse products and bad decisions
- What to do instead to be truly customer driven and deliver real customer value
With that in mind, let’s dive in…
Why Great Product Teams Avoid Feature Voting Tools like the Plague
There’s more than one way to build a great product, but there are a few traits that great product teams have in common:
- Data Informed: They measure the results of their work and use analytics and data to help them focus their efforts and see where they need to ask questions.
- Customer Driven: They think with the end user in mind, whether they understand it innately because they are it, or they deeply dig in to get to know and speak with end users of all types.
- Detail Oriented: The details are what separate okay and good from great when it comes to products. One of my favorite sites is dedicated to them: LittleBigDetails.com
- Product Sense: This obviously takes time to develop, but the best product teams include people who have and enforce having taste; they don’t fall for every fad, and bring a stamp to their work that people can see a bit of their fingerprints on.
- Collaborative: The Holy Grail of great product teams is being able to operate with the Cauldron approach that Steve Jobs used; everyone brings their best ideas, no one cares whose idea was what, and everyone focuses on creating the best possible solution.
These are all hard to build and can take time to develop. If you’re trying to level up a product org, you likely can only improve one at a time, but that’s a post for another day.
When it comes to feature voting, it works against all of these:
- Data Informed: As we’ll dig into more below, the data from feature voting is trash. Bad data is worse than no data.
- Customer Driven: It may seem like it brings some customer voice in, but feature voting is so distorted and warped, it represents a bastardized version of listening to and understanding your customer.
- Detail Oriented: When people vote on a feature, what are they really saying? You don’t know. You just know they clicked an up arrow, not the nuance of their needs.
- Product Sense: Feature voting tries to turn your customers into your product team. Don’t. do. that. You build a product team so your PM, designer, and engineers can create the best solution, which customers usually haven’t thought of.
- Collaborative: There’s nothing collaborative about a wall of feature votes on a screen. And without the real context of the problems to be solved, there’s no way to talk tradeoffs and iterate to something great.
Yet, despite all these being true, these are minor issues compared to the biggest issues with feature voting: They are fatally flawed from the start.
Let’s talk about why.
How Feature Voting Fails
Let’s deconstruct all the failed parts of using a feature voting system. It starts with a faulty foundation, and falls completely apart from there.
1) You get *features* not problems.
If you look at the average results of feature voting sites, what you’ll see is a lot of people asking for a feature. Typically posts say things like:
- Make a CSV export
- Build an integration with X
- Add tangential feature Y
Here’s the thing: Any of those requests should be the *start* of a conversation, not the answer by itself.
For instance, with a CSV export, all of these questions come to mind putting on my PM hat:
- What specifically do you want to export?
- Why do you need an export? How will you use the export?
- How would you like the export formatted?
- How often would you expect to need the export? Why that frequency?
All of these questions can lead you down a rabbit hole that realizes any of the following:
- A new report in your product would be better, and more up to date, than an export.
- The export is for a key weekly meeting for the customer, so automatically emailing the numbers would be even better.
- The export is only needed once a year, and needs to cover multiple sections of your product to really meet their needs.
The only way to get these kinds of insights is to talk directly to customers, and a feature voting site does not allow you to have those 1 on 1 conversations effectively.
2) People get easily sidetracked…
You’re using a product. Suddenly you notice something annoying, confusing, or missing. You decide you’ll share the feedback that their stacked bar chart that only has two shades of blue is very difficult to read and you’d like more control and granularity.
You notice they have a feedback button, or a link to their feature site and eagerly head over.
When you get there, you’re greeted by a list of dozens, if not hundreds of other options.Without meaning to, you start reading the other options, maybe clicking on a few. “Oh that sounds interesting…” you think.
Without meaning to, you start reading the other options, maybe clicking on a few. “Oh that sounds interesting…” you think.
But then you forget why you came, and either never get around to sharing the feedback you meant to, or giving only a partial explanation of your original thought.
Either way, the product team loses, as the feedback that motivated you to come to the page is much more valuable than a random upvote or two.
3) …Too many votes for something discourages future posting
While distraction is one problem, the sibling of it is people flat out giving up. If you look and see the top upvoted item was posted in 2017 and has 400 upvotes, what makes you think your new idea will ever get any attention? And do you think that anyone at the company is even really listening?
When feature voting apps were first getting popular about 5-10 years ago, I used to dig into them to try to get an idea of how product teams used them.
It was depressingly rare how often I’d see someone from the company actually active.
Even worse, when they were active, it was usually trying to explain why they’re not going to build the most popular items.
Now, feature voting sites can try to help with this by building some kind of algorithm to show “trending” or “most recent” items, but that’s really lipstick on a pig; this is just one of many significant problems.
(Fun aside: At a past job, I looked at our competitor’s UserVoice and noticed a number of posts from people asking for features we had that they didn’t. I passed this to our VP of Sales who then figured out who these people were and got some of them to switch to us.
So your Feature Voting page not only hurts your product, but makes it easier for your more ambitious competitors to steal your frustrated customers.)
4) Nobody wants yet another log in
One of the most famous stories in e-commerce history is the discovery that you can make *millions* more, and have double-digit improvements in checkout conversion by allowing people to check out as guests:
No one wants yet another log in, yet in many feature voting tools, the first thing you have to do is create another account to upvote or leave a suggestion.
As anyone who has worked on growth teams or on e-commerce conversion rates knows, adding steps to a process will always lead to more drop-off.
So let’s think about it…which is easier:
- Leave product you’re in to go to feature voting site
- Land on the overall page filled with existing suggestions
- Click on an item or to post your own feedback
- Sign up for an account (or if you’re a masochist, sign in with your account from last time)
- Find the way to add your suggestion and post it
Or, the direct way:
- Click on Intercom, a Feedback button, or link
- Send your message
If you’re measuring the rate of customers having feedback to actually sending, the latter option will far outperform the former.
But we’re getting ahead of ourselves on the solution, so let’s continue with the problems of Feature Voting.
5) Not all votes are equal.
Let’s look at two different Product Managers and see what they have:
PM 1: Talks to customers directly, gets feedback passed to them from other teams, and with help, organizes it all.
PM 2: Relies on the feature voting page, when convenient, to show how many people are asking for something like what they dreamed up.
PM 1 has 25 logged conversations where they or a colleague they trained asked some follow up questions for context, and understands this is “very important” not just a “nice to have.”
PM 2 has 50 upvotes for a feature that kinda sounds like what they’ve spec’d out.
Question: Who has more real data to back up their decision making?
- Answer: PM 1 in a landslide.
Upvotes != real customer feedback
The number of upvotes a Feature Voting submission has does not mean all of those people want the same thing.
Let me repeat that: The number of upvotes does not mean every upvote wants the same thing.
Here are some of the reasons someone may upvote it:
- “I want that feature exactly as described.”
- “Well, this has 25 votes already, and it’s close enough, I’ll upvote that instead of post mine.”
- “Oh that looks interesting, I wouldn’t mind that. Click.”
- “I think my coworker wanted that…”
- “I could have used that a few months ago” (and haven’t needed it since)
- “That sounds cool.”
All those upvotes and you don’t really have a quantitative count of customer input. Something could have 250 upvotes, but it’s the least important item. Or they could want different functionality or features as part of it.
You don’t know though, because all you have is an upvote, not a conversation, or even a few sentences from each of those people.
QUICK PM QUIZ:
You’re a SaaS PM that has a mix of SMB and mid-market customers. Which feature should you build:
- A) A feature that all 5 of your biggest customers say is critical to their workflow
- B) Dark Mode, because it’s highly upvoted on your Feature Voting tool
If you choose B, please think about a career change
6) Customers forget what they wanted if you wait to reach out
Do you remember what you were thinking a year ago? How about a month ago? A week?
For the vast majority of people, ideas, and feedback, are fleeting. When you’re in the moment doing something is the time you’re most in tune with the situation.
If I come back to you a year later asking, “Hey – I saw you upvoted Feature X last year. We’re finally working on it. What were you looking for?” Unless you have continued regularly experiencing the issue that prompted the vote, you’re unlikely to remember the request well.
You might have also changed jobs and be unreachable, gone on vacation when I reach out, or not even remember voting for something.
That means following up with all those upvoters when you finally get around to a feature, you’re unlikely to get nearly as many useful insights as you would in the moment.
Feedback is like milk…it goes bad quickly when raw.
That’s why it’s so important to talk to customers regularly, and ask them in the moment what the underlying problem is and why it matters to them. That’s when they’ll remember the context you need to truly understand their request.
Just because you can’t build something right away does not mean you can’t talk to customers about their needs and save it for later. You do that by talking to them (in chat, email, calls, etc), not by collecting votes and checking your Feature Voting app once a quarter for ideas.
From distractions to lost feedback, and murky data to blurred meaning, Feature Voting is fundamentally flawed from the start, and only gets worse the longer you use it.
Now, let’s talk about what to do instead.
What to do instead of Feature Voting:
By now you understand why it’s a terrible idea to add feature voting to your product, but that’s only half the battle.
You need to then focus on what to do instead.
1) Source feedback where it naturally occurs
Done right, your company can be a customer / feedback gathering machine. We did it when I was the 1st PM at KISSmetrics.
There are great sources of feedback available all around you:
- Sales Teams: They know what deals are closing and what deals they’re losing. And the best sales people know the difference between good losses (not a great fit) and bad losses (could have won and the customer would be happy).
- Customer Support: They deal with the angry, the frustrated, the annoyed, and the confused. Tap into their knowledge and fix their biggest problems…and you’ll fix your customers’ problems, too. They also will get feedback, so give them a way to pass it to you.
- Account Management: They’re in charge of keeping customers and making them successful, so they see the gap between the promises your sales team made, and the reality of using your product. Another gold mine of feedback and insights, and even a potential source of future junior PMs.
- UX Researchers: As they do usability testing and test new features, they likely hear customer gripes, questions, and feedback. Make sure that gets captured.
- Data Teams: Ever wanted to ask a specific subsegment of your audience a question? Your data team is your best friend in helping create all kinds of great segments to analyze, survey, and reach out to for customer interviews.
- You: If you’re a PM and not regularly talking to customers, what are you doing? Make the time, and build the habit both for features you’re actively thinking about and to generally get feedback.
Now, that sounds like a lot, and it is. Which is why you need to dig into one of the key, unheralded jobs of great product managers: Relationship Building.
Have Peer 1 on 1s with key people on other teams.
Choose some of the best people, or those most related to the part of your product you work on, and have peer 1 on 1s with those people in customer success, sales, and account management.
If you’re new to these, this post dives deep into how to have awesome peer 1 on 1s as a product manager.
As you meet with them every 4-8 weeks or so, remember to teach them how to fish! What this means is that you will:
- Explain to them your goals of gathering feedback and understanding customer needs they’re hearing, so they get what you’re trying to accomplish.
- Teach them how to ask a good followup question or two (“How important is this to you? What is most important to you in this request? Why?”) before passing information to you.
- Involve them in prioritization by telling them the threshold for passing it to you (i.e.- “Once you hear something 10 times, let me know” or “If it’s a customer >$N per year, pass immediately to me.”), thus keeping your signal to noise ratio strong.
You can also then share back with them how you’re listening and acting on what you hear from them.
Nothing puts a smile on the face of customer success like hearing “Yes, we’re finally fixing that bug you have been dealing with for months” or telling a sales person about a hot new feature you both know they can sell like crazy.
When you start sourcing information from all across your company, you’ll see that feature voting is so incredibly low quality compared to all the ways you can get detailed, context-filled, specific information from your peers.
2) Choose your focus
Feature Voting apps don’t know what your company’s priorities are this quarter, and neither do your customers. Yet, as far as a Feature Voting app knows, all requests are created equally.
That’s why it’s important for you to narrow your efforts as a PM by starting with a focus.
Some examples of proper focus would be:
- Retention: You need to improve retention and resurrection of old accounts for your social app.
- Churn: Your SaaS app is leaking too many customers, bringing a drag on growth.
- Growth: You need to create more organic growth through an improved viral coefficient.
- Conversion: Customers are adding to cart, but not buying. You need to find out why, because it is crushing your CAC.
- Activation: Why are customers signing up, but not completing the setup process?
- New Bookings: Your company is trying to go up market, and needs to identify the features that will allow for more medium sized business deals to close.
For each of these problems, a different segment of your customer base would be your target to talk to. Along with this, different problems and features would be most important to focus on.
By starting with your focus in mind, it narrows down your efforts significantly, and can help you ask the right questions of all your peers you’re now meeting with and talking to semi-regularly.
This context helps you make better decisions, and changes who you talk to and what you ask them.
3) Get Quantitative, too!
While ultimately the goal is to understand who your users are, how they use your product, the problems you solve (or need to solve in the future) and how you fit into their world, you must balance that qualitative information with quantitative data.
Being customer driven also means understanding the numerical side of your customer base. You likely have a few different personas and company types who use your product. You need to understand how that translates numerically to your business with answers to questions like:
- What % of our customer base is each type of business? How much of our total revenue do they correspond to?
- What customer types have the highest LTV?
- How do our core metrics compare when we slice our customer base by various properties (like company size, business type, various demographics, plan type, device, location/region, etc):
- Churn rate
- Rate of expansion
- Referral rate
- Conversion rate
And I’m sure you can think of many more. This kind of quantitative work is a priceless exercise, especially if you haven’t done it before.
When I was at KISSmetrics we dove deep into these and discovered that company size didn’t really matter, but when it came to business model, SaaS and Ecommerce businesses converted 2X as well, churned half as much, and thus had a much higher LTV.
Do you think that impacted our future product decisions?
Surveys are your friend, too.
Now, ideally you’d have infinite data you could easily query across your product to answer every question. However, that’s neither feasible, nor really desirable (it would be too costly, hard to maintain, etc)
For snapshots in time, and to jump-start efforts, surveys then become your friend.
This allows you to then take input from your customers (especially those interested enough to take a survey) and segment it based on the questions you ask.
That’s why for instance with my startup, Lighthouse, we ask people what department they work in. Certain departments convert better than others, while one department has proven to be a massive time waste. We automatically filter out the latter’s input, because we know it’s not useful.
However, surveys are not a guaranteed silver bullet. In fact, most people make a ton of mistakes using surveys by making them too long, asking too many open ended questions, or using confusing language.
That’s why I regularly revisit this awesome post from Profitwell on how to do product surveys.
The TLDR is simple:
- Keep it short
- Make as many questions multiple choice as possible
- Ask customers to mark the *most* and *least* important things instead of rating everything
Now, taking a step back, not only are you building an engine to gather all your feedback and to better understand your customers, but you should also be using your analytics and other quantitative data (sales & marketing numbers, conversions, custom queries from your data team, etc) to help prioritize what metrics you want to move.
And once again, this kind of data informed approach gives you much better information than the random, muddy data of feature votes. Here you’re understanding problems, not starting with features.
4) Make people feel heard!
I hinted at this above in the problems with Feature Voting section, and it bears repeating: your customers want to feel heard.
Posting or upvoting on a Feature Voting site is like the Suggestion Box in a Dilbert Comic:
The real way to make customers feel heard is for them to get a response from someone on the product team to things they ask for, and to occasionally see things they ask for fixed or added.
It could be its own post on how to do this, but for starters, here are 3 of my favorite approaches:
1) Tell your customers about new features launched
Any progress is good progress in the eyes of customers. It gives them hope you’ll get to some of their requests, and can really make their day when you finally build something they really wanted.
That’s why I’m a big advocate of sending semi-regular product update emails.
Doing so let’s them know you’re listening, gives you a place to thank those that gave feedback, and reduce churn as people recognize you’re improving the product regularly.
2) Make all your product emails have a real reply address so you can talk to people
I’m always stunned when companies send announcements and product emails and make them email@example.com. That’s a big missed opportunity.
Instead, make it a google group that sends to some of the product team. You’ll get fewer emails than you may be worried, but those that do really care. Your customers will appreciate being able to respond (often saying positive things and showing gratitude!) and with a simple, quick reply, you (or a coworker) can make them feel heard.
3) When you launch a feature based on feedback from a customer, email them personally.
How do you feel when you get personal notes from friends, family, or people you work with? Pretty good, right?
You can do the same for your customers simply by sending them a quick thank you note when their input is acted on.
If you’re a small startup, then this should be pretty easy. Sending 10 thank you’s to those that hopped on a call should take you 5 minutes.
As you scale, this can scale too. If need be, pull the names and emails of those that submitted feedback, had a flagged support ticket, did a usability test, etc and do a mail merge to send all of them a similar form note thanking them. Anything is better than nothing.
You can also enlist your coworkers, like for instance asking your Account Manager to reach out to their customers involved and let the AM share the good news.
This is win-win; it’s less work for you AND the Account Manager looks good to the customer, as it shows they can effectively pass along feedback that gets acted on by the product team.
Doing this not only makes people feel heard, but it also helps you build relationships that create power users and customer advisory boards. The more people feel heard, the more they’ll reach out and make your life easier as a product manager seeking out feedback, problems, and insights.
Isn’t this a lot of work?
Yes, this is a lot of work, but you’ll notice quite a bit of this is collaborative. That means you’re sharing the workload. And best of all, your hit rate on features built will go way up, so there’s less drama, more excitement, and you overall become more efficient.
All because you roll up your sleeves and do the work.
Take the time to be a product person who truly cares about their craft and builds processes and paths to directly learn from and speak with their customers. You’ll find that feature voting is then the last thing you’d want when you have all this direct, quality customer insight coming in.
Thanks to Ray Wu , Adil Majid, and Willis Jackson for feedback on this post.
Need help for you or your product team? I can be your coach.
I’m doing a limited number of engagements now helping with getting your first customers and early adopters, overcoming churn issues, and validating new features. You can sign up for a free intro call here, or learn more about what I do here.
Pingback: 10 Things I Love About Superhuman « Building Customer Driven SaaS Products | Jason Evanish