Record List to Excel processing very slowly

Good afternoon, I am working on a page that allows users to export large datasets as an Excel spreadsheet.   What I am doing is taking some inputs from the user and then running a SQL query that fetches the records in question.   That recordlist is then changed to a binary via the Record List to Excel step and then that binary is saved to a temporary table (which is purged nightly) to be fetched later when/if the user is ready to download the file.   

The issue I'm seeing is that the SQL query completes very quickly.   In the test I just ran I get back 550 rows and comes back in one second.   But when that record list hits the Record List to Excel step it processes there for over 45 seconds, which seems rather excessive.   Has anyone else experienced this?   Am I missing something obvious that would speed up this process?   The structure of the SiebelQuery has 28 elements  (or 28 columns returned)  so it's large but not a huge dataset.  

I’m not seeing how the timer would benefit me here.  The user input is necessary to determine the data in the excel file.   I would not be able to preemptively create the binaries so they are readily available on demand.

Josh, probably to find the root cause you need to revisit the columns being fetched from the SQL and deselect the ones which are likely to have lots of data.

Regarding timer approach, this is not a solution to make your process faster but would help to improve user experience. Assuming you only need user interaction to get some inputs and then user downloads the file later but not immediately, the user inputs can be taken and stored on some database table which would act as timer queue. Then the timer based on set schedule would pick an item and perform the rest of the steps for you.

I see your point, our users all have email associated so I could just set a timer to run every 15 minutes and to fetch the data and send the file via email.

Most of the columns are varchar(50) but there is one comment field that is quite long.   I'll try deselecting that first to see if the conversion to binary happens more quickly.