俄中选委正式宣布普京赢得总统选举

In case this is not already planned, the second change about replacing 0 with Vote should not apply to posts you own or are otherwise ineligible to vote on. Them being zero score should still be made clear to the post author.

For Q&A that has no score or answers that have equal up and down votes the zero will be replaced with the word “vote”.

This seems unnecessarily confusing. I imagine a fair few users would assume the vote count is locked, or hidden, or requires something extra (privilege, payment), or something.

Keeping the zero but having some extra notice around it to incentivise voting would probably have been clearer.

Another alternative I have seen on GOG.com is that when there are no user submitted scores for an item yet, the aggregate rating shows up as "Be first to rate":

No. No. Please oh please, no. No.

No.

I had no idea about this until it happened to me today and I had to use Google to find this meta post.

I've been a member here for 10+ years and vote all the time. I just counted 29 times in the since the beginning of Dec with approx 60% downvotes. I don't understand why I am being targeted in this experiment but I hate it.

The word "vote" feels like a command - which in the context of what the OP has described it basically is - and it reminds me of those signs in the movie They Live that say "Obey" everywhere.

Please, for the love of all things, stop this madness. Or at the very least give me an option of permanently opting out.

I don't like resorting to this but using stylus I reverted the madness like this:

div.js-vote-count[data-value="0"] > span {
    visibility: hidden;
    font-size: 0!important;
}
div.js-vote-count[data-value="0"] > span::before {
    visibility: visible;
    content: '0';
    display: inline-block;
    font-size: var(--fs-body3);
}

^{(Selectors updated 2025-08-07)}

I want to address a point that JNat made in a comment:

The experiment will focus on upvoting only; answers only; on the top answer only, if there are multiple. The pop-over will not show concurrently to the currently existing banners. We will see what we learn, and go from there.

This feels like a terrible idea to me. Upvotes and downvotes are really important, and the Question here makes it very apparent that downvotes aren't in use any where near as much as upvotes, with only 6% and 20% using their votes respectively. Downvotes are the clear loser in that scenario. This is evidenced in the site analytics as well:

Only putting this on the "Top answer" feels even worse to me, it feels like you are promoting that specific answer; the top upvoted answer is NOT always the best answer. Pile on votes for old, bad answers, are a problem enough, and a lack of downvotes on those answers is a problem as well; don't make the problem worse by suggesting people add more pile on upvotes to the "top" answer when it could be far from the best. The pop-up should be post agnostic, and vote type agnostic as well, in my opinion.

If you won't implement this, will you be able to determine after the prompt how many people voted on an answer different to the one they were prompted to, and/or if they downvoted (on the answer you suggested to upvote)?

Further in regards to Downvotes, and the lack of them, I wrote a "quick" SEDE to see the weight of use of downvotes to upvotes. I included only uses that are (currently) about to cast both up and down votes, and have voted at least 100 times. I was not expecting the sheer volume of people that have never cast a downvote but cast 100 upvotes; 39432 users (averaging over 300 votes each). There are more users that have used 100% of their votes for upvotes that users who have less <90% of their votes for upvotes.

These users could be making a lot of upvotes, however, rarely (never) downvote, and this is really important. I therefore really suggest that Stack Overflow reconsider weighing on Upvotes alone; ask people to vote that's all you need to do. How they vote, up or down, is less important; we need people to be using their votes correctly, and just upvoting doesn't do that.

Result of the SEDE:

And visually if you are so inclined

Change the text from "Did this work?" to "Is this useful?" in the new voting popover:

• Not every answer is a matter of making something work. For example, understanding questions (as opposed to problem-solving questions). (Thanks to starball for pointing this out in comments.)
• This is consistent with the current text on mouseover: "This answer is useful".
• The question "Did this work?" suggests that the user should test the answer before voting.
  • Testing the answer is not necessary for the upvote. I often upvote answers from which I learn something potentially useful, without necessarily testing whether the answer actually works.
  • Saying "it is useful" instead of "it works" lowers the "psychological bar" to vote on an answer.
  • The converse is also true. For example, a duplicate answer (which is redundant, and contributes nothing new, given an already existing answer), may very well work, but it is not useful.

Having lived a day with this applying to my profile, I really don't like it. Maybe I'm in the minority (and I'm sure I'm not really the target audience, with how frequently I vote), but I really don't like having to think an extra step about what score a post is at when I view it. I spent the better part of my morning this morning writing a userscript to hide it, and it was frustratingly less successful than usual because of the duct tape method this test was implemented (it appears to be applied after the page is done painting, via JavaScript, rather than properly implemented via server side logic).

A much better method would be to show the score at 0 like normal and focus instead on just displaying the modal that encourages users to vote more prominently/frequently when they haven't voted in a while.

If people feel strongly enough to have an opinion, they don't need to be reminded that voting exists; It's right there at the top of every post in the most prominent location on the page. We should instead be focusing on understanding why people aren't voting rather than using cheap gimmicks to annoy people into voting.

At best this feature would encourage a half dozen people 4 times a year to cast votes. There are far more obvious places we could look to find people who currently aren't voting but would be more than happy to.

If you are using UBLock or Brave, you can remove it with this rule:

stackoverflow.com##+js(rpnt, #Text, /^Vote$/, 0)

As far as I'm aware, this is the only place where a single text node with the value "Vote" exists on the site, but if someone has a more restrictive way to limit this to avoid affecting other areas in the future, update it.

ref

I was looking forward to this experiment and I had optimistic attitude towards it... until I actually got this A/B test.

It's awful. It's distracting. It's confusing. It's a visual noise. Just NO!

Maybe consider some polite tooltip drawing attention to the buttons instead. Replacing the post score with a word "VOTE" was a horrific idea. I knew about this experiment and I still got confused multiple times about what's going on. It does NOT look like a nudge towards voting.

status-declined

Talking about replacing score 0 with "Vote": Targeting a zero score answer has no point in it. If anything, you want to target a zero vote answer. It is meaningful to ask for zero vote answers to be evaluated, but not zero score ones.

An answer could be heavily voted on with with 12 upvotes and 12 downvotes, and you would still display 'vote'? This makes no sense.

A zero total score is as worth as any other number, and not so special/ugly to be targeted.

We don’t see many users who hold voting privileges participating in voting.

This breaks down as follows, on a monthly average:

• Only 20% of registered users who have the upvote privilege are using it.
• Only 6% of registered users who have the downvoting privilege are using it.

These stats on their own do not support the claim that many people who could vote do not vote. The sprinkle of statistics might make the post seem "data-driven" but the intelligent reader will see through.

Can you clarify if it's 20% of all registered users who have the upvote privilege? Or only those that were active in a particular month who don't ever vote?

If it's all, then it could just be that 20% of registered users don't ever visit SO.

If it's active users, it could still be that they just visit one question that month, and don't vote on 1 question. I wouldn't call that a problem.

Even prolific voters don't vote on every question/answer they look at and that's probably also not a goal.

What's the ideal rate of voting per question view in your view? How many people achieve it and how many don't? Those are the sorts of things that might make this more data driven.

What's the distribution of question views without voting for those 20% that don't vote? That would be interesting insight. If someone visits 100 questions in a month and doesn't vote then, yeah, they should be encouraged to do so. But you haven't shown that with your data above.

Regarding the update to extend this experiment to all users on January 16th, here is my feedback

I hate the Vote prompt.

From the "question"

Why we are doing this

We don’t see many users who hold voting privileges participating in voting. This breaks down as follows, on a monthly average:

• Only 20% of registered users who have the upvote privilege are using it.
• Only 6% of registered users who have the downvoting privilege are using it.

Voting is one of the most significant actions users can take to signal quality on the network. We would like to see what minor actions we can take to encourage users to vote more frequently on content they visit. These experiments will only be visible to those with voting privileges.

I might be among the users who rarely upvote, but that is because very few posts deserve an upvote. It's improbable that I will vote more frequently due to the Vote prompt.

Related

• How do I opt out or manually revert the new "vote encouragement" experiment?

For Q&A that has no score or answers that have equal up and down votes the zero will be replaced with the word “vote”.

Since you seem to have ignored the request to remove this entirely, could you at least give users the option to disable this and/or auto-remove it for users who have cast a certain number of votes.

Note: My later words in this answer post were written before the clarification that downvotes will not be encouraged as part of this experiment, or at least this round of this experiment. I like the intent, but am starting to not like the implementation.

I'm disappointed to hear that this will only encourage upvoting (and only on the top answer post) and not downvoting, considering that downvoting was mentioned explicitly in the question post.

I'm also turning around on the word "Vote" replacing zero-score indicator. I thought I wouldn't mind it, and I thought it would read informationally rather than like a command, but now that it's on my screen, I do mind it, and it does read like a command, which is off-putting.

(original answer):

I support this. I'd love to see people make use of their voting privilege more (provided they're voting based on whether things are useful to them or not). I suspect that not voting is a habit formed from votes appearing to be meaningless before gaining privileges (and to a large extent, that's true).

Questions-

• When they have downvote privileges, what will the popover look like? Will it only encourage them to vote up? Voting down is equally important.
• What about questions? The example screenshot shows the popover for an answer post, but what about question posts?
• What happens when there are multiple answer posts?
• I like that the popover is close to the buttons themselves, but given the two questions above, was it considered to use the existing stuff like the "welcome back- remember to vote" and "you haven't voted on questions/answers in a while" things, which appear at the top of the page?

This morning, we expanded the vote prompt and popover to all users.

Does that mean the experiment has graduated?

Or does that mean that the experiment has moved from A/B testing to only B testing?

It is horrible. Is there a Tampermonkey script to get rid of the abomination?

For a long time user of SO, this prompt is terrible.

The vote count is one of the most crucial pieces of information in the UI and you've just decided to obscure it.

When I'm looking for an answer, I usually first look for the most voted one, read it, cross-reference it with a few more top answers and then skim the rest, focusing more on the ones with low vote count that may be new and target some latest updates. Notice how important to this process is to have fast and reliable way to check the answer score.

I've just got targeted by the A/B test and it was like my brain instantly hitting the wall.

The previously simple brain procedure of:

> parse_int(vote_of_answer_0)
0
> parse_int(vote_of_answer_1)
0

turned into:

> parse_int(vote_of_answer_0)
Exception: number format is incorrect
> parse_int(vote_of_answer_0)
Exception: number format is incorrect
> parse_int(vote_of_answer_0)
Exception: Operation timed out!
Switching off fast track thinking model.
Engaging logic engine and redirecting full attention to the number.
Exception: The data is not in an expected format. Expected "int" but got "string".
Inference: the layout has been changed (70% probability)
Searching for the vote count in a different location...
No matches found.
Checking memory banks...
Found a match: "Voting experiment to encourage people who rarely vote to upvote".
Inference: "vote" means "0" (50% probability due to poor quality of the recalled memory and possible changes after that announcement)
Exception: the information is not sufficient to proceed as normal
Calculating the next step...
Clicking on the "vote" to validate the score...
0/0
Inference: "vote" mean "0" (95% probability)
Updating the fast track model...
Warning: updating automatic processes usually takes a long time and multiple attempts - the results may be unreliable
Disengaging attention.
Trying to restore fast track thinking model.
> parse_int(vote_of_answer_1)
Exception: number format is incorrect
> parse_int(vote_of_answer_1)
"vote" interpreted as "0"
Exception: inconsistency found - "vote" is much too long to be "0"
Critical error: the fast track thinking may be compromised - forcing inconsistency resolution mode.
Engaging full attention.
Retrieving the last working solution for this problem...
Clicking on the "vote" to validate the score...
0/0
Updating the fast track model...
Warning: updating automatic processes usually takes a long time and multiple attempts - the results may be unreliable
Disengaging attention.

I might get used to "vote" after some time but even then it will require more attention to parse than just a plain number. And even if not, till then you'll manage to waste a lot of my time. Now multiply this by the number of users who will get similar problems as me, which is most of the long-time SO users and ponder if those experiments of yours are really worth doing if everyone is telling you that this is a bad idea.

The idea to promote voting is good,but the second experiment "vote prompt" is heavily distracting, asking us the reason for this abnormality, and many are eventually distracted and moved to this meta question, than focusing on the answers there. This experiment doesn't support the voting much theoretically than other alternatives like popover.

I've been seeing this for a while now and had to seek this post out to see the background of this change. I still participate in voting, but nowhere near as much. I've voted recently and it's still showing for me.

If you want me to vote more, you're just going to get more downvotes across python with this. "Vote" is a very evocative word on its own and, accepting my own fallibility, it's just going to make me downvote questions where I would otherwise give the benefit of the doubt and assume there might be some fringe knowledge of some library that I don't understand/seem unable to research.

The term "Vote" as a seeming command massively trumps the "New Contributor" flag in my head. It's not nice to admit that but it's no different than colouring a notification in red to compel a person to read it IMO. In this case, it is definitely detrimental for the outcome you want from me.

Has this now been released to everybody? Or what changed in the experiment? As of maybe an hour or two (not sure when exactly it started) I now get the "Vote" text. I was not in the "Vote" group earlier today on this same machine. I see several more people reporting they see "Vote" as of extremely recently.

A bit of real-user feedback: I had no idea that "Vote" was supposed to be an encouragement to vote: I initially thought the vote counts were broken.

To make it a bit more obvious, you could potentially update the tooltip? It still says "View upvote and downvote totals" (I'm over the threshold to see this prompt). Given that the posts in question will always have a score of 0, I suggest changing it to explain the text somewhat. Something like this (a modified version of /help/why-vote):

You haven't voted on this post yet. Voting is central to our model of providing quality questions and answers; good content rises up and incorrect content falls to the bottom. Learn More...

TL;DR - I see two big flaws of how this experiment was implemented:

1. People who are active voters receive the experience - if the goal of the experiment is to remind people to vote, then the users who are regular voters should not receive the experience.
2. Once I vote (multiple times BTW), I still receive the experience. Once I voted, the reminder should appear once a sensible amount of time has passed (days or weeks), instead of appearing the next time I visit another post (which in my case was seconds after I voted).

I just ran into this on a random answer to a random question - neither the question nor the other answers were showing this:

I have voted a lot since I joined SO, so please don't show this to people that are active voters.

The heck, I voted a few times this evening, and now this "Vote" thing started to appear on even more posts. Looks like a bug, why am I reminded to vote a few seconds after I casted votes on other posts?

I've become a very infrequent user of Stack Exchange, and when I saw the text "vote" in place of the score, I thought this was some sort of bug—that the actual voting score didn't load and a placeholder text was put there instead.

For the life of me, I wouldn't have expected this behavior, and I used to be very active on this site.

I haven't posted answers on Meta Stack Overflow yet, but this time I feel I have to post an answer to this.

First, a score of zero on posts is replaced by "Vote", which is significantly wider than the vote arrow buttons, this alone breaks the layout and makes it ugly in my opinion.

Second, there are two vote buttons, upvote and downvote. "Vote" doesn't specify which type of vote to cast, do you ask people to downvote? If you want people to upvote, you need to make it scream "Upvote" instead of just "Vote", which uglifies the layout even more.

Third, if you want to encourage people to vote, why do you only want to encourage people to upvote? Why the bias against downvotes? Downvotes help control content quality on Stack Exchange network. If we only cast upvotes, then low quality posts will be out of control, and this network will cease to be a high-quality repository of information.

Last, what is the criteria for upvoting?

This question shows research effort; it is useful and clear

Is any of these metrics considered when that number zero is replaced with "Vote"? Did you consider the quality of the post when you remind people to vote?

There are lots of people posting questions with code generated by AI, questions without code, questions that blatantly ask people to write code for them, questions that are unclear, and completely off-topic questions...

Are you asking people to upvote those questions? Is this your latest attempt to turn this network into yet another social media?

Our test group showed a 3.8% increase in new monthly voters.

That annoys thousands of actual heavy SO users.

What exactly do you mean by an increase in "monthly" voters if this experiment only lasted about a week?

I seem to see this on every 0-score question, along with the console error:

Unable to load nonexistent manifest entry 'recent-posts'. Check that your file path is correct.

Is this currently broken?

We plan to prompt users who have voting privileges but have not engaged with the feature much (e.g., once every three months, never voted before, etc.).

What's the exact criteria you're testing? The current ones are very vaguely specified as a non-exhaustive list of:

• once every three months (did you mean haven't voted in past 3 months? Otherwise over which period do you average?)
• Never voted before (redundant category: it's included in the above)
• etc. (who else is included?)

This feature is annoying to me. But luckily, it seems that I can still opt-out this feature by logout StackOverflow, so maybe not too bad(?).

The number of voting is an important information to the questions and answers. As they make a perfect guide to me about the quality of the answer before I really read it. So, to my browsing pattern, I would simply see the word "Vote" even before I had read anything about the content. So I would wonder how can I vote without reading the content.

Before you encourage voting, you should be clear about why people (up) vote answers. Maybe it is...

1. Because the answer is helpful to me;
2. Because I regular click the vote button before I read anything;
3. Because the vote (like) button is large and round;
4. Because there is a button I can touch on;

And I believe the answer is not (1) under your consideration.

If the popover is dismissed, the user won’t see the prompt for a period of time.

What period of time? Maybe you should just not show it if someone has dismissed N, e.g. 3, times in a row, otherwise you're causing popover-mania for those who simply don't want to vote.

An alternative is to use exponential backoff. E.g. show it again after a day, then 2 days, then 4 days, then 8 days, ... so that one only has to dismiss too often.

I really don't like the idea that this increases the incumbent top answer's advantage by being shown only for the first answer. Pile on votes are already a problem as people scroll till they find a solution that works which leads to recently submitted better answers having a big disadvantage re voting.

I get that you don't want to show the pop-over on all questions and answers. But you could show it on a randomly selected answer instead.

On pageload, you decide which answer number to show it on and that's where it's going to show up.