Work Problem Involving Messy Data. Attempting to Create "Standard" Time for Widgets

Here is what I am trying to do and why this is throwing me off. It just feels like there is a way to do this mathematically, but I just don't have the knowledge to come up with it.

At my work we have been doing a time study to determine what the best time to set as "standard" time for each widget. The associates have been manually tracking how long it takes them to complete each widget with seven months of data.

Since we get new employees that require training and we have employees that are experienced and fast, I wanted to try to find a "fair" number. To do this I figured that removing the outliers would be the best course of action, so I decided that a trimmed mean would give me a good average time while removing a certain percentage of the fastest and slowest times. I would also calculate the standard deviation for each function and set a cutoff for if the widget was considered "volatile" and we would just continue to track manually. Everything was going to be awesome!

This is where the problem happened.

In the tool that the associates are using to track their times combines like items for the same day. If they did 17 widgets for the day, I just get 17 widgets and the total minutes. Each individual widget isn't counted. My plan was ruined! It does break it down by associate and by day. There are many instances where an associate may only complete one of a certain task in a day, but that obviously happen all the time.

My question for you is given this slop of data I have, how could I calculate the appropriate average time for each widget using some method to remove the fastest and the slowest and how would could I determine the volatility for each job function?


Ambassador to the humans
I'm running a bit slow today but I'm having a hard time understanding what your data looks like. Can you provide a sample?
So the data come across like this:

This is only for one day. This associate had 3 items for function 1 and a total of 15 minutes.

This is part of the time in minutes and number of items. There are around 3500 lines. there are about 25 functions.
I think I might be stuck with just taking the average, but that seems like such a terrible way to find a "good" time.