So many choices, so much time.
Six place trigonometry tables to Kepler?
A good set of navigation charts to Eric the Red?
Translate the SAS survival manual into Celtic? Oh, no, wait, they didn’t have a written language.
The further back you go, the harder it is to provide a book that anyone would recognize as a source of important information. Give the Picts a book on archery, they might have thrown it at the Celts, but they certainly wouldn’t have used to make bows and arrows.
The information in a basic practical nursing course would revolutionize almost any Paleolithic tribe, but writing is not going to give them the information, and even having it falls a bit short of believing and using it.
Latin texts would be readable in much of Europe, over a very long time, but the reception of the information is still a highly variable thing. Translating the Handbook of Chemistry and Physics into Latin is probably a whole thread.
A good basic electronics text, sent to Benjamin Franklin is an interesting thought, though. He had the type of mind that would use it for a starting point, and slog on through for years. You wouldn’t have to translate it, although he might have some trouble with modern American English.
Just saying “a book about germ theory” isn’t really enough. Which book? A twenty-first century text about infectious disease doesn’t spend much time on the concept of microbe existence, or the methods of identifying microorganisms. In the time of Pasteur, there were damned few people to whom it would not be incomprehensible.
I suppose if you could ignore the translation problem altogether, a basic high school hygiene course delivered to almost any small early tribal culture could be the most effective in achieving major change, but it would be almost impossible to predict the nature of the change.
They Boy Scout Manual could have made a difference, in a lot of times and places.
Tris