The web is missing an enormous amount of what people actually know. Not because those people are unqualified — because there was never a way in.
Picture the person who has watched the same river gauge most mornings for thirty years. Not for a job — because that is what you do when you live there. They can tell you what that river does before a flood in a way no hydrology page will, because the thing they know is not a formula, it is a pattern they have watched. Search for the river and you will get a rafting outfitter, a real-estate listing, and a page of aggregated weather data that is wrong about the low-water crossing.
They are not absent from the results because they lack authority. They are more authoritative than everything on that page. They are absent because publishing on the web requires a site, an audience, and a tolerance for being optimised against — and none of those has the slightest thing to do with knowing something.
That person is a composite. You probably know one anyway — nearly every reader of this can name someone whose knowledge is not on the web, and that recognition is itself the argument.
The machinist who can tell you which of the six answers about hardening that steel will crack it, and why the popular one keeps getting repeated.
The midwife with twenty-two years of practice, watching a search result recommend something her profession stopped doing in 2011.
The county archivist who knows that the answer to your question is in a bound volume on the third shelf, and that no one has ever put it online.
The birder with four decades of local records, who can date a migration shift you will not find published anywhere.
The retired engineer who wrote the manual everyone is now quoting badly.
None of these people is going to start a blog. Several of them tried, once. What they have in common is not a lack of expertise or a lack of generosity — it is that every existing route asks them for the wrong thing.
Write an encyclopedia entry and defend it against strangers. Build an audience first, then be believed. Post an answer and accept a score. Learn what a search engine rewards and shape your knowledge to fit it. Each of these asks for a skill that has nothing to do with the knowledge, and filters for people who have that second skill instead.
The web did not select for expertise. It selected for the willingness to publish.
Those are very different populations, and we have spent thirty years mistaking one for the other.
Here is the part that makes this tractable rather than merely sad.
The knowledge is usually not locked in someone's head. It is in a forum thread from 2009 that answers the question completely and ranks nowhere. It is in a trade journal that nothing indexes. It is in a society newsletter, a mailing-list archive, a manual, a bound volume on the third shelf. It exists, in writing, and it is unfindable.
Which means the useful thing an expert can do is not write. It is point.
They already know where the good material is. That knowledge — this source is right, that one is repeating an error, the real answer is in this paragraph — is the exact judgement that ranking is missing, and it is the one thing no crawler can produce. A machine can tell you a page exists. Only someone who knows the subject can tell you it is correct.
Not write what you know. Not build an audience. Not defend it in a comment section.
You already know where the good stuff is. Point at it.
A source, a passage, and what it answers. A minute's work, checkable by anyone, and it changes what the next person finds.
There is an honest problem with asking someone to improve the web: the web is enormous, and one person's contribution vanishes into it. Landing on the first page of a general search is not something an individual can do by being right. That is a discouraging thing to discover after you have helped.
So we do not start there. We start inside a herd — a group formed around a subject, holding its own slice of the index. Within it, the material you point at is not competing with the whole web. It is competing with what that group has found, which is a contest that rewards knowing the subject.
You can be first in a herd about hill-country rivers. You will never be first for rivers. And the first is the one that matters to the twelve people who needed it.
Herds also solve something quieter. Nobody has to decide which subjects are important — no editor picks the categories, and we do not promote one. A herd exists because people cared enough to form it. The index bends toward what its members actually know, which is the only bias worth having.
The Oxford English Dictionary took seventy years, and for most of that time it was visibly incomplete. It published in instalments from 1884 and was not done until 1928 — useful for forty-four years while everyone could see the gaps. Nobody thought this was a failure. It was a project, and projects have a beginning, a middle and an end.
The web is treated as though it simply is — a natural feature that arrived complete, whose contents are a fact rather than a consequence. It is not. It is a corpus about forty years into being assembled, missing enormous regions, and currently being assembled by whoever benefits from the traffic.
That is a direction, not a destiny. Dictionaries had the same problem and it was solved by strangers who read carefully and sent in what they found.
If you have read this far and thought of the thing you know, that is the thing. Not a general contribution — the specific correction you have been carrying around, the source you keep having to send people, the answer that is right there in a document nobody has indexed.
Bring one. Find the herd for it, or start one. The whole apparatus — the checks, the ranking, the index — exists to make sure that when you point at something true, it stays pointed at.