I just posted to our blog about our Third Quarter Community Roadmap for 2020.
Please see the post for details on the projects that our Community and Public Platform teams aim to work on this quarter.
We are interested in your thoughts and questions regarding our upcoming projects and priorities on the roadmap.
Update: Y'all should really be looking at how GitHub does this: https://github.com/github/roadmap
Just a brilliant bit of information design. Each task follows a consistent format:
Textual and visual metadata is then used to indicate the progress of each task and what it pertains to. The resulting roadmap can be quickly scanned and parsed for relevant information. I cannot recommend this highly enough!
Nineteen tasks for this quarter, eh? That's... Ambitious...
I published my thoughts on the last two quarters' tasks a bit ago, and while they sparked some good discussion at least one person got pretty salty about some of the assumptions I'd made about the nature of the tasks. So in the interest of transparency, I figured I'd post my assumptions up front this time.
This is how I've interpreted the tasks laid out for the current quarter:
An employee clarified in the comments that this was completed earlier this month.
Assumed test for success: Platform is re-used for D&I stuff mentioned in later task.
Hooray! Please publish results. Surveys are like the Doomsday Machine in that Dr. Strangelove movie: they're only effective if folks know about them. I've seen what happens when you do regular surveys and don't publish the results or do anything in response to them: folks stop taking the survey.
Assumed test for success: results are published somewhere.
This sounds useful. Also useful: giving them something useful to do.
Assumed test for success: moderator council figures out why they exist.
This is done-ish, unless too many moderators refuse to sign it. In which case, you'd better hope that the "Election Automation" task further down moves from "discovery" to "implementation" very quickly.
Assumed test for success: there are still enough moderators come September.
Good. I'm epically tardy with responding to letters, but 10 months is impressive even to me.
Assumed test for success: a response emerges that isn't, "there will be a response".
This is already underway!
Assumed test for success: fewer repeat suspensions, same number of people reviewing (or more), meta is not attacked by mobs of reviewers with torches and hay forks upset about their suspensions.
So... Subject matter experts are going to do something, somewhere?
Assumed test for success: subject matter experts do something, somewhere.
Another survey! Gonna just re-use what I wrote for the last one:
Hooray! Please publish results. Surveys are like the Doomsday Machine in that Dr. Strangelove movie: they're only effective if folks know about them. I've seen what happens when you do regular surveys and don't publish the results or do anything in response to them: folks stop taking the survey.
Assumed test for success: results are published somewhere.
This is good. There was a LOT of feedback, both on meta and spread across the 'Net. Looking forward to an epic report here - don't forget that you can ask a dev to increase the maximum post length on meta!
Assumed test for success: a report is published that covers the outcome of the test in detail, summarizes all the responses, and lays out next steps.
This is a good idea. But... It was a good idea a year ago too - actually finding or creating useful training turned out to be much more difficult. A big part of the problem was that moderators don't stand to benefit much from the standard "D&I" courses that most new employees have to sit through these days (y'know. The ones where an animation slowly explains that fraud and harassment are bad, bad things) - what they actually need is the training that the folks giving D&I training get. And that turned out to be harder to line up than a Flash animation.
Assumed test for success: moderators report receiving training that is actually useful and relevant to what they do as moderators.
The lessons of the last attempt here should be:
Those are... some pretty low bars to clear, but... They should've been the first time too.
Assumed test for success: nobody reports receiving irrelevant, insulting, or scary emails.
I hope this means the international versions of Stack Overflow are going to get more support.
Assumed test for success: a new era of first-class support for languages and cultures other than English / Western. Failing that, Aki gets access to CM tools so that she can finally just edit the Japanese help center articles.
I first asked for this in 2012. If it actually gets done before 2022, I'll cheer.
Assumed test for success: literally anything happens with election tooling.
I assume this means the TL is still sitting in that hack of a room I put together on MSE after certain people did the textual equivalent of throwing Molotov cocktails into the old TL, repeatedly, for weeks. Sad. I hate to say it, but this isn't going to work unless there are TL moderation tools for moderators as well - there are only four CMs left, and while they probably can still manage 24/7 coverage if need-be... It's sort of cruel to ask for that. It was pretty cruel already last year.
Assumed test for success: next time someone sets off a bomb in the TL, it doesn't burn for days.
Complicated title for something that there's no description of. I am going to guess... Another survey?
My copypasta key is wearing out; please see notes under "Moderator Quarterly Survey" and "Downvotes Research".
Big, if true.
Assumed test for success: any sort of plan or analysis of Area 51 is published. Ideally one that doesn't just badly reinvent something Robert proposed 6 years ago.
Please, please, please just steal one from a site that already has a good Markdown editor. SO's editor was best-in-class 12 years ago, but it's been lapped many times now; heck, https://stackedit.io started with SO's editor and has long ago left it in the dust. Y'all don't have to try for something world-changing here, just aim for parity with what's already winning: take the one from GitHub, or StackEdit, or Visual Studio Code, or innumerable other places that've actually spent serious time making good editors over the past decade. Then use the time you save to beef up support for pasting in code - that's the big failure-scenario for new SO users.
P.S. Please don't steal Reddit's; it sucks.
Assumed test for success: an editor that rivals the one in Visual Studio Code. Or is the one in Visual Studio Code.
Well, that sounds vaguely ominous. But, legal things generally do.
Assumed test for success: Stack Overflow complies with all relevant European laws regarding the privacy and retention of personal information.
That's it! If there are errors or inaccuracies in any of the above, feel free to edit, or just post corrections to the blog.
I for incomplete, because nobody tried to burn the place down again. (Or less optimistically -- that might not be a criteria that can be evaluated in the quarter the assignment is completed... might need another metric by which to judge success).
The blog post implies that "Downvotes Research" is going to be limited to Stack Overflow here. I suspect you'll find very different results if you also start looking at the more subjective sites instead of just the technically oriented ones, and especially the sites which deal with more controversial topics (e.g. IPS, Politics or any of the religion sites). And that's not even getting into the whole "Voting is different on Meta" effect.
Also, will you be looking into how users feel after being downvoted, or are you just focusing on why users cast downvotes for now?
A question after the alt stuff (and how Mith ended up doing it in the end) - you speak of inclusivity training. What areas in particular?
(The alt stuff, or rather, non-descriptive alt or info only available in low-qual pictures, hurts anybody with vision impairment. That's why it's important. If you're interested, I'm publishing an article about a11y/inclusivity in the field of gaming centered on a blind user tomorrow. Their perspective/take on what is available out there is more enlightening than reading about it through a third-party source)
I'd just like to take the time to give a quick thank you to Yaakov and the team for a quick response to some of the concerns raised about this blog post.
In Discord, a concern was raised about how D&I training was mentioned for moderators, but there was no mention of that same training for staff. The blog post was edited to address this fairly quickly.
There was also the image alt text issue brought up in Seb's answer, which ended up with the alt text that I added here being put into the blog post itself.
Sara's comments about planning on doing better in the future were also appreciated.
So thanks for the quick responses and addressing the feedback!
Since I'm in the middle of trying to get some moderation in place on the moderator end of things, and rustling up some missing functionality - does this involve retrofitting moderation tools into the "legacy" TL, or enhancing the temp-permanent one on the MSE chat-server?
If it's the latter - would it be possible to take into account tools/abilities for TL Room Owners in the process?
If there's any secret squirrel stuff going on - I'm happy to move the conversation elsewhere.
Regarding the "Iterating on educational emails (September)" item, isn't it generally too late to tell a new user how the site works after they've signed up and asked their first question?
Hopefully, the new user has read the site's tour page, but that lacks some essential knowledge, such as emphasising that downvotes are not a personal attack, and that the user needs to tend to their question for perhaps a couple of hours after posting it, which could help with the previous item, "Downvotes research (July)."
Some sites have asked to have their vote to close threshold lowered, from 5 to 3.
It's something that would be really nice, and while some CMs had access to do it in the past, now they do not. As it requires a dev to help our CM, could this be on the roadmap? It's maybe not a big item, but some sites are waiting for such change.
It's a change that was planned just before some CM firings, but now our remaining CMs have that task left in their hands
I ask, as I keep poking our CM for that in TL, but they have no dev time allowed for the task. It's not their fault, but if the item is not on the roadmap then that does not help them to get the resources to do it.
Your efforts are very much appreciated!
I find it extremely encouraging that you managed to get onto a path that seems to combine very different interests: On the one hand, the legitimate business need to attract users and paying customers. On the other hand, such core elements like improving moderation culture and tooling. And while diversity is still in, your plan acknowledges that "user happiness" also has a lot to do with "technical aspects should need to be logical and working"!
It is great to see how you folks kept course during this global crisis, and I am looking forward to participating with increasing frequency in the coming months, simply because you created an environment that invites your users to do so!
Downvotes research (July)
Receiving downvotes on Stack Overflow can be a frustrating and confusing experience. We currently ask users to downvote posts that are not useful or are unclear, but this can be subjective and interpreted in different ways. We will run a short targeted onsite survey to better understand what motivates users to downvote a post and use this data to inform inclusion and engagement opportunities.
Whom do you plan to inform here?
From inclusion and engagement part it reads like you want to engage and inform users that downvote posts.
If that is true, I think you are missing main point. Downvotes are for poor and unsuitable content, they are cast because users asking questions (or posting answers) are not following site guidelines.
If you eliminate users that will not follow guidelines no matter what, what remains are users that were not properly informed about guidelines before posting. This is huge pain point, especially on Stack Overflow. You need to improve informing about guidelines and informing about consequences before users start writing their first posts.
