Incremental load is used to save time while doing a reload by not reloading all historical extracted data that you already have gathered one more time. To implement incremental load you will have to store the historical data updated with new each time you do the reload. The historical data you store in this way is much faster to merge/concatenate with the newly extracted data from your source - hence you will potentially save a lot of time by doing incremental loads.
The downside of using it is that it requires some logic implemented in a load script and you have to maintain your historical data and make sure of their integrity. Luckily there are some best practices explaining this both in general data warehousing terms but also special technical whitepapers or technical briefs from Qlik about this in a Qlik Sense / QlikView setting.
