I'm in the process of building some XML that will subsequently be transformed by XLST into HTML in the form of tables.

Am I correct in saying that the XML needs to be well structured to efficiently use XSLT for the transformation?

For example, I should be structuring the XML such that it flows into a table row by row using the for-each construct right?

Something like this:
<root>
  <row>
    <columns>
    <columns>
    ...

etc.

Yes, the more the structure of the data is explicit in the XML tagging, the easier it will be to process it in XSLT. For example it's much easier to process

<section>
  <title/><p/><p/><p/>
</section>
<section>
  <title/><p/><p/><p/>
</section>

than

<title/><p/><p/><p/><title/><p/><p/><p/>

Michael Kay
http://www.saxonica.com/
Author, XSLT Programmer's Reference and XPath 2.0 Programmer's Reference

Thanks - appreciated.