Uh, the STL does not make the sales projections. Corporate makes those. If, however, the STL thinks the projections are going to be way off (say the nearby university is going back to school a month earlier because they switched from quarters to semesters), I think they can ask for it to be adjusted.
I don't know exactly how sales converts into payroll hours, but I bet it's fairly complicated and sucks (mostly because it gives so few hours, but that's more profitable so that's what they're going to do).
However, I do know how the flex calculator works. Basically, if you do really well or really badly (in sales) during the month, you gain or lose payroll for that month. Each month starts fresh. There is a cushion of 3%; if your sales are within 3% (above or below) your forecasted sales, then you do not gain or lose any hours. After that, you gain/lose the same percentage of payroll that you have done in sales.
For example, if you do 2.5% better than your forecasted sales, you will not gain any payroll. But if you do 4% better than forecast, you will earn 1% of your payroll. If your monthly payroll is 10,000 hours, then you will have gained 100 hours. If you do 3.01% better than forecast, you gain .01% of your payroll. If you do 10% worse than forecast, you will lose 7% of your payroll.
This means that if your store starts out by missing sales by 5% for the first week of the month, your STL might start slashing hours. That could be a trend that continues and makes you lose tons and tons of hours by the end of the month. Plus, one bad snowstorm (like some parts of the country are seeing now) can easily make you miss sales by tens of thousands of dollars, or potentially even close the store for the day (closing might actually be better because you'd save the payroll, I guess). A store could be doing 2.90% worse than forecast going into the last Saturday of the month (just barely making enough to maintain payroll) and then miss that Saturday by $30,000 and end up missing payroll for the month.