Wikipedia:Consensus venue

Breadth of consensus is heavily dependent on venue and process
and (later) its outcome. Consensus breadth is the extent to which it represents the community's view rather than just that of some special interest. As a matter of clear policy (at ), a broader consensus always trumps a narrower one.

A matter affecting a large number of articles or editors should be resolved in the most obvious site-wide, not isolated, place for that discussion, or the claimed consensus for it is weak and suspect.

The more significant it is and/or the more likely it is to raise controversy, the more it should be neutrally "advertised", e.g. with an template, at WP:Village pump (policy) (WP:VPPOL), via WP:Centralized discussion, at talk pages of implicated policies and guidelines, at relevant wikiprojects, etc., and centralized to the appropriate main policy/guideline talk pages like WP:AT, WT:N, or WT:MOS not those of sub-guidelines.

If consensus is reached, the change should be written into the main page that governs that sort of thing (or the appropriate well-accepted sub-page thereof), whether it be a policy, a guideline, or even a "near-guideline" essay like WP:Arguments to avoid in deletion discussions or WP:Bold, revert, discuss, and regardless of the subject (a guideline on reliable sources is not different in this regard from one on style or editor behavior). A new would-be rule buried in some wikiproject advice-page essay is largely going to be ignored in any dispute that arises, and for good reason.

Backwater discussion = dubious consensus
It is typically enough to just widely advertise some discussion on a wikiproject page, even if RfC were used to attempt to direct attention there. Editors routinely ignore, as parochial noise, any insular proposals to change policy (when such RfCs are even clear that's the goal, which is rarely). A "rule" arrived at this way will be lost to everyone but the most esoteric collectors of wikibureaucracy, not even findable in the talk archives of the policy/guideline in question.

All that will happen as the result of a parochial, echo-chamber consensus discussion is some pages over-controlled by the wikiproject will start to diverge from the actual site-wide rules, other editors will eventually notice and make it conform again, a dispute will break out, and the wikiproject will falsely claim a consensus to have made up their own rules. the broad discussion will happen, tempers will already be hot by then, and the false consensus will be overturned, but after the erosion of considerable editorial goodwill. This pattern has played out again and again.

While sometimes well-advertised RfCs on more provincial pages can work, if broadly advertised to WP:VPPOL, to related policy and guideline talk pages, to topical noticeboards, to other wikiprojects with an interest in the scope, etc., this is, especially for non-trivial matters. VPPOL in particular is kind of a firehose of poor ideas, and few people read it with much seriousness any longer. This is even more true of WP:Village pump (proposals), the dumping ground of every other perennial bad idea on Wikipedia. Village pump discussions rarely gain any traction at all unless hosted directly in VPPOL. So, if you think you're going to redefine what "reliable sources" means for football player bios by having a wikiproject discussion, you are sadly mistaken. That's probably a good enough venue for, say, the background color of player infoboxes, but no discussion held there is going to be considered broad enough for major policy and guideline changes.

If a wikiproject (which is just a page at which some editors are coordinating how they want to collaborate, nothing more) comes up with a "rule" that conflicts with site-wide policies/guidelines or with normal English writing, their isolated little consensus among themselves, mostly representing a uniform viewpoint, is just the beginning of the consensus-building process.

A small group of editors can't change policy just by agreeing amongst themselves
A group of editors with a peeve to pet who fail to engage in this baseline level of process is worse than a false consensus, but an anti-consensus faction, improperly trying to invoke the "ignore all rules" (WP:IAR) principle without the justification IAR requires. IAR is the last resort, process has failed, and only when invoking it will make the encyclopedia objectively better (not just subjectively better from a particular viewpoint).

The process for changing any policypage is to discuss changing it on its talk page or at VPPOL, until resolution is reached on whether the proposal is helpful or not.

Even if a wikiproject has hundreds of active editors who came to some conclusion that other editors may disagree with, that is still just a special interest agreeing with itself that it's a special interest and defining how special it is; it is not a Wikipedia consensus. Even when their would-be rule does not conflict with site-wide rules, they can't reasonably expect the rest of the encyclopedia to follow their lead if they won't seek a broader consensus, outside their own talk page. Too often it actually will conflict with site-wide rules. Wikiprojects are also not hive minds; it is never actually the case that everyone in an active wikiproject is in perfect agreement; individuals who act as if their personal opinion represents a whole wikiprojects' views are either self-delusional or trying to game the system.

When wikiprojects, reference desks, portals, collaborations, or other limited groups of Wikipedians are engaged in any kind of "we are our own sovereign wikination" false-consensus pattern, it is not helping Wikipedia in any way.