The hardest part isn’t writing something that millions of people find critically important—it’s ensuring that society will continue to find the same things important for centuries. A temporal value-alignment problem.
which is a hard task. and irrelevant to the OP's question.
There are many things that people would want to preserve, but is of no importance to the grand scheme of humanity. You're basically saying that unless your works is of such importance that millions of people generations to come would voluntarily preserve it, the works do not deserve preservation.
I'm not quite following why this is not relevant to the OP's question. I suppose I'm saying that it's not just a hard task, but an almost impossible one. Lots of people (2000 years ago, say) thought that what they were writing was of world-historical significance; today, we barely know their names. Others dashed off dictated letters to small, fragile communities of early followers and just happened to have created among the most well preserved texts of all time.
All I'm really saying here is that among the titanic forces determining which is which, the type of ink used seems not to be a matter of much importance.
What does last? Making something a tradition can help smooth over logic. Making it part of a long running organisation would help (organisations can exist for multiple centuries).
Perhaps an organisation dedicated to preserving messages through the ages, and encourage - say - a Christmas message? More people involved mean more care.
Sure, you have the biggest chance of having something preserved if enough people care about it. This is a case where, as you rightly point out, technology matters not.
But we can still discuss technologies of preservation, especially in the light of non-physical creations like works presented on a digital machine.
I think this is actually wrong. It certainly doesn't hurt to write on fired clay tablets, or to "store" your papyrus documents in an extremely low-humidity environment. That might enable accidental preservation. But at the of the day (or at the end of the millennium), things survive because lots of people care about them.
You mention Anglo Saxon. That's an interesting example, because the entire corpus of Anglo Saxon literature is easily anthologized in a single volume, and there is exactly one extant copy of what most scholars regard as the most important thing written in the language (Beowulf). But the truth is this: Most of it is gone, because people stopped caring about what was written in it.
But psalters? Bibles? Church histories? It's an embarrassment of riches. Why? Because people cared very deeply about those things.
And translation helps, but again, it's not quite enough. The instructions for my dehumidifier are in eight different languages. Where did I put that again?
To come at it from another angle: Finding a paper copy of a 1955 Seattle phonebook is extremely difficult. But why? It was written on paper, there were thousands and thousands of copies printed, none of them have actually disintegrated into dust, and yet . . .
So: My advice would be to make that web page pertain to something of epochal importance to millions of people. They'll do the rest.