Warning: severely biased Go (amateur) programmer here! As such, my first reaction was 'oh no! another fantastic resource going away!'
But on the other hand, I asked myself: how long ago did I ask anything on the Go Collective, participated in whatever thread there was, or replied/edited/closed any unanswered question there?
And I realised that the answer was 'quite some time ago'. Since I'm constantly browsing the Web in search for answers to the weirdest questions that an amateur programmer always has, it would be a good sign if Google/Bing would consistently show answers coming from the Go Collective.
This is hardly the case. Perhaps, because as @blackgreen mentioned, there is so much documentation about Go, spread across the whole 'net.
The main hurdle I personally face is to seek, among the potential questions/answers, those that have some genuine content worth reading. This is what is supposed to happen 'automatically' as participants vote up and down on them. But if a large proportion of 'noise' is mixed in the middle of some 'signal', it's easier to find worthless content and vote it up (just because it's easy to find!) than promoting a well-thought question up, which is so deeply buried that it hardly gets any readers, so... it will never be seen/read, and never been upvoted.
Note that this is not an issue specific to the Go Collective, or even of the constellation of sites around StackOverflow; all user-moderated content that relies exclusively on voting — from YouTube videos to corporate feature request voting on applications or even games — suffer from this. It's unavoidable, especially if there is a lot of user-moderated content. It's even hard to predict what kind of content will 'float to the top' — the concept of virality (the tipping point which makes certain content 'suddenly' to be seen — and voted! — by an exponentially-growing number of users in a short time) is, at the time of writing, completely unexplained (and unexplainable). If it weren't, marketeers would be using it all the time :-) It's something even not at the grasp of ML or other forms of AI — 'profiling virality', as of today, hasn't resulted in a simple set of rules to follow that will produce virality.
That said, I believe that the main issue with anything like the Go Collective, user-moderated mostly through voting content up and down, is very unlikely to produce so-called 'good content' of knowledge-base-quality consistently, and, perhaps more significantly for SO, automatically. This is, naturally enough, just my perception — you could call it a conjecture at best, i.e. something that can be tested with the scientific method in some way (and possibly refute it), but without explaining what the method should be.
In other words: the current model of SO does not automatically guarantee the production of a consistent amount of 'good content'. In fact, it might be argued that, on the long term, as the number of people using SO grows, the signal-to-noise ratio worsens, simply because the number of votes to keep 'good content' at the top will require to increase as well. I'm not a well-trained statistician, and forgot most of my university maths, but it seems to me that this is a universal trend of any system that is based solely on voting on 'good content'.
The major reason, of course, is that curated content is moderated by specialists, while user-generated upvoted content is merely a 'bikini contest': people voting up or down an answer are just manifesting its usefulness for themselves, not necessarily evaluating the answer's overall content quality. While there is a correlation between 'good content' and 'upvoting' — or else the whole of SO/SE would be completely worthless — such a correlation is not immediately obvious, and I guess that the SO/SE sysadmins, looking at their logs, will know very well the ratio between the number of times an answer is viewed and the actual number of voters. Or, if you prefer, there is no incentive for upvoting an answer that solves a problem and is well-written, unless you're a stakeholder (you wish a particular answer from a particularly knowledgeable user to be at the top of the searches of your community, to make sure that your community is valuable for all).
This is, indeed, the problem with anything that relies on collaborative volunteering (a.k.a. crowdsourcing). It's hard to offer any meaningful incentive for participants, except those incentives that participants (volunteers) develop for themselves. Once that self-incentivation is removed (through external reasons, e.g. rules change, original developers leave, excessive advertising starting to bother, cyberbullying, whatever...), it's harder and harder to keep such volunteers around.
Sorry about the convoluted way I'm answering this topic. Again, my opinion is not substantiated by scientific evidence. I'm just addressing the philosophical issue here: I believe (and aye, it's just a belief) that by merely creating a Collective, or some other sort of 'grouping' (under different assumptions and rules), is not guaranteed to automatically produce good content and an engaging community. It may happen, but it will be hit-and-miss. The only thing that can be consistently predicted is that, as a site/group/thread/forum/Collective increases the number of participants, the signal-to-noise ratio drops — something we've seen happening everywhere, from the earliest Bulletin-Board Systems to, well, USENET News.
In fact, the main reason usually stated for the complete abandoning of USENET News (except as a relay mechanism for pirated content, and as a niche for some die-hard users) — and, in a sense, IRC, although less so — has been shown to be the complete lack of moderator control. The move to 'corporate' websites essentially replicating USENET News have added value: since they're privately owned, their owners can, if they want, specify the terms and conditions under which their service is run; they can, if they wish, moderate content (and have the appropriate legal framework set in the terms & conditions that allow them to do so), expel users, and so forth. They even can, at a whim, change all the rules of engagement — if they prove not to be adequate for their original intent. This was impossible under USENET News, and the result was that so-called 'quality content & knowledge' migrated elsewhere (since the volunteers producing such content lost their incentive to 'stay' on USENET News).
The way each company operates their system varies, and it's not clear to me what approaches work, and what doesn't. Medium — and now Quora — seem to believe that their 'best' content producers should be placed behind a paywall, to encourage them to produce content only for 'premium' users. Again, in my opinion, this is not a good long-term guarantee of survival, mostly because if someone is willing to pay for good content, then it's much more likely they'll prefer to pay for professionally curated content, such as the content produced by, say, newspapers, or scientific journals, most of which provide high-quality content, at least high enough to get willing subscribers to the service. Accessing 'slightly better-written content' behind the paywalls of Quora or Medium is, well... not for me to say, but I expect that both companies (and possibly many others) will not thrive in the long term (disclaimer: both have contacted me via their sales department to become a writer behind their paywall — I declined, for the simple reason that I firmly believe that my content is not worthy to be 'paid for', even if just symbolically so, especially because I offer it for free on my own blog and, naturally enough, scattered among a plethora of websites out there). The main problem with them is that there is a 'free' level, with less and less good content producers (which will be co-opted, over time, to write-for-pay), where content will eventually be so bad that free users of the service will only have bad experiences there and therefore not believe it's worth paying for eventually 'slightly better' content...
Now, I understand that the ultimate goal for the Collectives would be to have an official sponsor as a source of revenue for Stack Exchange Inc. But that requires a bit more than merely setting up a technical way of 'building' such a specialised team/group/collective. I understand that you guys have put a reasonable amount in time & money to develop such technical functionalities, and now you naturally expect to get a return on that investment.
Here is my very simple solution, which will — I believe — be deemed acceptable by many SO 'hard core' users: get sponsors to sponsor tags addressing their products/services. Sponsors would have their logo appearing on the sidebar — e.g. not 'disturbing' the free flow of content on the Q&A themselves — but such sponsorship would not be merely 'advertising', but rather a kind of 'sponsored endorsement'. In other words: Google may sponsor the #go tag, because Go was originally developed at Google (and is still mostly maintained by them). Microsoft may sponsor #.NET, or C#. Apple may sponsor the whole of the thinkdifferent site, or optionally may sponsor tags such as #apple, #mac, or #iphone. Naturally enough, even non-profits may sponsor (in some way deemed reasonable for Stack Exchange Inc.) their tags as well, e.g. #firefox sponsored by the Mozilla Foundation, and so forth. In fact, in the non-technical forums, it might be even easier to get sponsors — from universities sponsoring tags in the science forums to entertainment giants (Disney sponsoring content related to Star Wars, or HBO for content related to the Fire & Ice saga in its many incarnations...).
While this limits, to an extent, the number of potential sponsors, at least such sponsors would be curated by Stack Exchange Inc., which would require of those sponsors (as you did with Google and the Go Collective) that they have a reasonable cause for interest in specific tags. But this might nt always be the case, and it would be up to Stack Exchange Inc. to decide what is 'reasonable', and what is not. For instance, it may be unavoidable that Oracle sponsors the #mysql tag; but it might be harder to decide if Bloomberg (aye, the financial news giant) should be allowed to sponsor the #sqlite tag, just because they are members of the SQLite Consortium (that sponsors the ongoing development of SQLite)! And nobody should be allowed to sponsor #sql, since it's hard to prove that that the specifications of the SQL language (not the server-specific dialects!) are 'owned' by anyone.
I'm not claiming that this would be a 'perfect' solution. But if your issue is how to measure the quality of sponsored content (so that such sponsors may be happy with the results), I don't believe that having 'collectives', under the current model, or a different-but-related set of rules, is the way to get you guys a regular stream of income — simply because you don't control the quality of the content, and, as such, you cannot give such metrics to your potential sponsors, neither can you make any claims regarding the 'overall quality' of content. That's impossible under any system operating under 'bikini contest' rules.
I'm aware that one might argue that Wikipedia, for example, 'works', in the sense it has high quality, fully user-generated content, and continues to grow and be relevant, in spite of an ever-growing community of users. Wikipedia therefore seems to go against the 'universal rule' that the signal-to-noise ratio goes down as the number of users grow exponentially.
However, Wikipedia does not work as a 'bikini contest'. Instead, it has several layers of curation. The so-called 'Wiktators', at their many levels, and numbered in the thousands (I've seen claims — wthout proof — that there are be more than 100,000 active Wiktators, at the many levels), make sure that content retains at least some semblance of quality — at the very least, content with mediocre quality will be flagged as such (alerting the potential visitor about the lack of accuracy, lack of reliable sources, and so forth). But such flagging does not happen automatically via voting up or voting down content! Rather, a human curator will read the whole article and make any necessary changes (including its deletion!). Such changes will then be reviewed by a higher-level moderator, and so forth — ultimately, Jimbo will rule as a court of last appeal (a role which he has refused to accept, and, as far as I could read about it, he never needed to exercise this 'ultimate power' in the past two decades).
It's conceivable that SO & friends may, at some point, develop into a similar knowledge base (albeit under a completely different model), but I would think that this will not really happen, for the simple reason that the 'ranking' of articles is mostly decided by voting in the bikini contest model, and only secondarily by its moderators. Since the ratio of moderators to participants is very low, it means that the few volunteers that patiently work through the questions & answers every day will constantly be overwhelmed with the sheer amount of data to process — for free. They will do their best effort to keep up, but there is a limit to how much they can do.
The main reason for that is perhaps that the rules for writing good questions and good answers are merely guidelines, which may be enforced a posteriori when moderators review content. But, until that happens, the 'bad content' remains there, indexed by search engines and web crawlers, taking votes up and down, and potentially allowing others, inspired by the 'bad content', to write similarly 'bad' questions & answers. Eventually, a moderator may step in and remove that content, or flag it for review at a higher level, etc., but since all of that takes precious volunteer time, it may not be effective except on the smallest communities.
Although technically Wikipedia is constructed 'from the bottom up' (that's what Jimbo originally intended), the reality is that it is moderated 'from the top down'. Aye, anyone can edit any article... but that doesn't mean that such an article will stay. As soon as someone writes/changes and article, they know that a moderator has just been flagged; and those moderators will make changes, according to an ever-increasing set of rules, which they do enforce, ruthlessly. That doesn't mean that people are discouraged to contribute to Wikipedia: it's just made very clear that either you abide by 'their' rules, or your content will be deleted, period. You may appeal, of course, but you really need to have very strong reasons, substantiated by the rules themselves, in order to have a chance to ever get heard. As such, naturally enough, only those willing to go through such a process of heavy editing, close moderation and curation, and abide by a huge set of rules, will actually spend their precious time to contribute — but they will so in a form and manner that promotes 'good' content (from the Wikipedia's moderators' perspective of what good content is — and they can define it objectively to a reasonably high degree!).
I don't think that you can promote good, quality content by merely waving a wand and allowing people to vote articles up and down. Rather, I believe that such content naturally emerges and stands out from the crowd because Google is more likely to find it. If you wish, it's thanks to the increasingly complex algorithms of search engines (possibly including SO's own!) that 'good content', like the proverbial needle, pops out from the (hay)stack under the force of a powerful magnet.
Stack Exchange Inc has managed to attract an impressive number of excellent writers (both on the 'answers' part as well as on the 'questions' part!) who managed — 'collectively' (pardon the pun!) — to generate a very decent amount of content over the many, many years. The better that content, the more people view it; the higher it ranks on search engines (not necessarily because it has a massive amount of votes!); which, in turn, attracts even more viewers, and so on. Even if I cannot estimate the number of queries that come from search engines, and the number that come from SO & friends own search, my guess is that the vast majority of viewers will basically have a question and thus search for some answer, expecting to find a list of 'good' matches; since this is usually the case (I'm talking about my own experience here!), viewers come back over and over again. Being therefore a good reference (and one where permalinks don't just stop working after a few years — aye, I consider that 'feature' critical to success!), the Stack Exchange Network becomes a de facto place to direct people to — no matter if it's on Super User, on Stack Overflow, on Unix & Linux, on Stack Exchange, or wherever Google finds a match. Ultimately, the site itself is less important than the content that is written there. And that's why artificially promoting a 'best of the best answers' — either through Collectives or any such similar grouping — will do little to enhance the overall experience. Personally, if I wish to get an answer to a Go question, I don't really worry if that answer is specifically inside the Go Collective, or outside it (maybe because someone 'forgot' to tag it adequately). I'm also not looking every day at the list of the last questions from the Go Collective and go through them sequentially, just to learn something new; the strictly Q&A format is not compelling for such a usage (especially because Q&As can have any possible origin — they're not necessarily grouped together to 'make sense'): the Community Wiki (being, well, a Wiki...) is (arguably) better for that; in fact, you might be more successful at getting sponsors for the Community Wiki(s)! (Not having used it in quite a long while, I admit I don't really know how popular it is these days...)
That said... I challenge you to prove me wrong on all points 😉
