Josiah knows that formatting changes to the first row of a list automatically clues Excel to the fact that the row is a header row so it won’t be included in sorting. He wonders, though, what exactly Excel looks for when it determines that the first row of data is a header row. For example, just making the first row bold does not seem to be enough, so he wonders what it really takes.
Microsoft has not published the algorithm it uses to make this determination, but one can intuit the formatting differences that are necessary. A good indicator seems to be if the data in the first row is different than the data in subsequent rows. For instance, if a column contains dates, times, or numbers (and are formatted as dates, times, or numeric formats), but the first cell in the column contains text (and aren’t formatted as a date, time, or number), then Excel can hazard a guess that the first row contains headers.
Another flag for Excel seems to be if your header row has borders turned on, whereas the rest of your cells do not. (I make it a point to always make the first row bold with a bottom border, and this seems to do the trick every time.)
You should note, however, that Excel doesn’t do well with multi-row headers or blank cells in a single-row header. It only really recognizes, automatically, a single header row. If you have a multi-row header, select the data you want sorted before you sort it. If your header row contains blank cells, then Excel (for some reason) interprets the row as part of your data. Again, you’ll want to select the actual rows to be sorted before actually sorting.