I'm not sure I completely agree. Citations should indicate which ideas are not yours and provide pointers to the original sources of those ideas and one should strive to make sure they do that. However, they're not a magic incantation that needs to accompany every non-original thought and walls of citations are often difficult to read.
Consider a paragraph like this*:
Johns, Jacobs, Jingleheimer, and Schmidt (2015, p. 15) state that "the
weather is going to be good today", a claim supported by several other
experts as well. (Jones, 2015; Xu, 2015b p. 9). However, we argue
claims like "the weather is going to be good today" are meaningless in
New England, where the conditions change rapidly and unexpectedly.
Relying on a claim about how the 'weather is going to be today' for the entire day may cause the listener all sorts of discomfort.
Repeating the entire 50-character citation** three times in three sentences would not add much value. The phrasing and punctuation clearly indicate what Jones and colleagues thought and, at the first mention, the citation is right there if the reader wants to follow up on it. However, there are two important caveats.
Books--and even papers--aren't necessarily read from beginning to end. To ensure that the goals of the citation are achieved, you need to ensure that anyone reading the uncited or "informally-cited" parts encounters the citation too. People are unlikely to pluck a sentence from the middle of a paragraph, but they might skip from section to section or paragraph to paragraph, so the citation should be repeated anew in each section/paragraph/figure legend. You may have a bit more latitude if the text is clearly a critique of one specific work: if the document is entitled "A critique of 'X, Y, and Z, 1995'", don't pepper every other sentence
with the full citation***.
Citing something for one fact/idea/quote also doesn't "absolve" you of the need to cite them for a second fact/idea/quote, even from the same document. The reader may not know that Johns and colleagues also made predictions about the stock market and the fifth race at Belmont.
In summary, make sure that it's clear which ideas are your own and which are borrowed, and indicate where the borrowed ideas may be found. Once you've done that, make the text as readable as you can.
* Example cribbed from a now-closed question on the same topic.
** Some citation styles do have a "short form" for subsequent mentions of the same work. That certainly helps here too.
*** Though you may want to be more liberal if the repeated citations include useful information (e.g., a page number).