When saving a Word document, at least in Word 2003, you have the option of saving as "Web Page" or "Web Page, Filtered". You should get a smaller file if you use the filtered web page option. E.g., for one particular Word document, I found the size was half as much when I used the filtered option versus the unfiltered option, i.e. 26 KB for the filtered file versus 54 KB for the unfiltered version.
If you select "Web Page, Filtered" for the output file format, Word doesn't include tags that only have meaning to itself. Those tags might be useful if you are reopening the file to be edited again with Word, but don't need to be there for people viewing the document in their web browser or if it is to be edited later with an HTML editor. E.g., for one document the following code was in the head section of the HTML file in the unfiltered version, but not the filtered version.
<!--[if gte mso 9]><xml>
<o:DocumentProperties>
<o:Author>Gail V. Williams</o:Author>
<o:LastAuthor>John Smith</o:LastAuthor>
<o:Revision>2</o:Revision>
<o:TotalTime>2</o:TotalTime>
<o:LastPrinted>2008-10-22T10:25:00Z</o:LastPrinted>
<o:Created>2008-12-05T21:57:00Z</o:Created>
<o:LastSaved>2008-12-05T21:57:00Z</o:LastSaved>
<o:Pages>1</o:Pages>
<o:Words>1108</o:Words>
<o:Characters>6317</o:Characters>
<o:Company>Home</o:Company>
<o:Lines>52</o:Lines>
<o:Paragraphs>14</o:Paragraphs>
<o:CharactersWithSpaces>7411</o:CharactersWithSpaces>
<o:Version>11.9999</o:Version>
</o:DocumentProperties>
</xml><!--[if gte mso 9]><xml>When you save a document as a filtered webpage, you will get a warning such as "Saving Test.doc in this format (Web Page, Filtered) will remove Office-specific tags. Some Office features may not be available when you reopen this page. Do you want to save the document in this format?". If you retain the original Word document in .doc format as well as the new filtered HTML version of the file, you can always re-edit the original copy, if you have any concerns about needing to retain the Microsoft Word specific information.
References:
