Western popular music is often structured in a series of nested structures; canonically:
Also, "measure" is really a notational convention; in some songs, measures can absolutely be key structural elements, but in other songs they really don't impact listening/analysis. But, because measures are so important to traditional notation, song structure diagrams are built primarily from measures.
Let's use Michael Jackson's "Beat It" as our primary text. My song structure diagram for it is:
Nothing about this is automatic; it is the product of a number of subjective choices I made, and there are many viable alternatives, as we'll see.
Time passes in reading order, from top-to-bottom, left-to-right, so the first measure is this:
except due to the low resolution image used it actually looks like this:
The measures are subdivided primarily left-to-right into beats, typically quarter notes; the three vertical stripes creating four blocks side-by-side, and then are further subdivided vertically into two eighth notes. (This final subdivision violates the top-to-bottom left-to-right rule, but it is the least important structural element in the analysis, existing primarily to allow distinguishing between simple and compound meters—whether there are two or three vertical subdivisions—and to allow unusual time signatures to be notated when they arise. But the intent is for them to be mostly ignored, the measure treated as the primary unit.)
"Beat It" is an example of a song where measures are not important to structural analysis, as, after the intro, everything uses 2-bar phrases.
The main "Beat It" riff is two measures long. This is clear in the guitar phrase itself, and also in the structure of repetition in the drum pattern. The riff is repeated twice in a call & response format, where the difference between the two is whether the last note of the riff is played once or played twice.
These two levels of phrasing can be indicated in a few ways in song structure diagrams, or can even be ignored.
There are three primary tools (and a rare fourth) used in song structure diagrams to help understand phrases:
Color will be discussed in the final section.
Finally, "long measures" is a device I use upon occasion when a song already features mixed meters. In the same way that it is sometimes better to notate using "6/4" rather than splitting something arbitrarily into "4/4 + 2/4", I will use long measures to indicate when a song features an extended phrase without any clear measure breaks. Two examples of this which feature chord changes every two beats are the dramatic punctuating verse phrases in "Truffles" by Wheatus, and the ambiguous choruses of "Wish You Well" by Katie Herzig.

In each case, these "measures" lack any internal emphasis that suggests where they should naturally be subdivided. And, again, I usually only do this if the song already requires mixed meter notation; for example, in "Truffles", some of these long phrases are an odd number of beats, requiring mixed meters for notation.
In most song structure diagrams, color is the primary indicator, and section labels are only used the first time each section appears, to avoid overwhelming the diagram with text. Because "Beat It" is particularly straightforward and unambiguous, I considered it a reasonable tradeoff to include all the section labels.
For example, in "Beat It", the final row of the intro uses the chorus red color, because that section of the music introduces the chorus riff exactly as it sounds during the chorus. It doesn't make sense to label it as the chorus, though.
Each section gets its own gross hues (e.g. chorus pink-red, verse teal), but there are additional smaller color changes used in a variety of ways. In the verses, the fifth and sixth measures have a different chord, while all the others are identical, so their color is altered to make that change visible. But this isn't something I always use color for. In general a single row in a single color can represent many sounds and different chords during that row, but typically means the same thing repeated on successive rows.
A common rule is that if two rows are the same color—or the same pattern of multiple colors—then they are "the same" musically (although they could have varying arrangements). But this is not universal; on songs without much internal repetition in the verse or chorus, sections may be multiple rows of a single color throughout those changes, because the color just demarcates the section.
In "Beat It", the guitar solo is given a color variation even though it is musically identical to the verse music, because plainly it is structurally its own section of the song. (This is an arbitrary choice, and I might have used the verse color on another day.)
Here are three examples of colors taken to extreme:
![]() | ![]() | ![]() |