Hi essdee,
I cannot provide a % difference between your browser render and LoadComplete replay timings, but can provide the following tips and insight:
First, if you'd like to understand the various LoadComplete elements that affect execution time, please reference the following article: https://support.smartbear.com/viewarticle/64933/
There are a few places where wait times and pauses come into play. You can change think time between pages by manually editing the recorded values in the scenario editor, which it sounds like you have done already. Alternatively, you can change these values at the test level. This link will provide more details: https://support.smartbear.com/viewarticle/64847/
Lastly, you may have some recorded pauses between individual requests, which happens depending on how your browser handles requests in parallel connections. You can view and change these via the Parallel Request editor. Note that changing these timings may negatively affect your scenario execution if, for example, the requests are sent out of order. You should only edit these timings if you have some unexpectedly long pauses between requests, which can happen with asynchronous requests, for example. More detail is provided here (scroll down a bit): https://support.smartbear.com/viewarticle/64930/ and documentation on the Parallel Request editor is available here: https://support.smartbear.com/viewarticle/65144/
Hope this helps!