Retail Price Change Non-Linear Regression Analysis

In my retail company we sell a range of food and non-food items.
The non-food items need to be ordered 12 months in advance from our suppliers and also hold our highest margin over our core food business.

At the moment we have far too much stock in our business which is crippling our operations. I am keen to cut down on these volumes but do not want to do so by the wrong amount that will cripple our profits. We can only keep 12 weeks worth of stock on sale before we need to clear all items to zero which costs a significant amount of money. We do not price change any items between 0-4 weeks of coming on sale.

The information I have is as follows:
Average Sell through in week 1 is 42%
Average Sell through in week 2 is 18%
Average Sell through in week 3 is 9%
Average Sell through in week 4 is 5%
The rest of the stock is reduced between weeks 5-12

Our year to date price change is 1.01% of total sales and the category share of non food items is 9.42% of the total shop. The margin we have for non food items is 17%.

The break even point is a 1.6% price change and whilst the year to date price change is only 1.01% of total sales I am confident that most of this price change occurs past the 4 weeks (when no price changes occur) meaning we are holding a majority of the price change in the last 26% of stock volumes.

N.B. We only have enough space to merchandised 4 weeks worth of stock before we have to reduce it to make space for new stock.

Currently the True Margin (after price change) is 6%. How can I maximise this to be closer to the 17% potential it has.