The hire packet has 67 pages in it, with some single-page and some multi-page forms.  Each form has its own document classification with an OCR first page identification condition.  The zones vary for each doc class, but are typically in the top third of the page. The first page of each packet has a bar code with the employee number in it, which is stored in a token collector so it can be applied to each doc class with a token retriever.

The goal is to speed up the scanning process by skipping pages that have very small text, such as the secondary pages on the W4 and the State Tax Withholding forms.

While I'm not sure what your identification conditions are, so this might not be relevant, if you put the OCR process in Page Processing, you can configure which pages will be OCRed (and which will be skipped).  There's two ways to do this:

1) Use the Page Range property

2) Use a Conditional process

Since it sounds like you always want to skip Page 2 of certain document classes, you can probably just use the Page Range property.  It allows somewhat complex patterns, which are documented here.  So, for example, if you always want to skip Page 2, you could use the pattern: 1, 3-1000 (1000 is just an arbitrarily large number to make sure it doesn't unintentionally skip pages before the end of the document).

If you had a more complex situation than just skipping a specific page (e.g., Page 2), you could use a Conditional process with the OCR process inside of it.