Normalizing cumulative data

I have a vast amount of sales and repair data for a range of products, and I would like to compare repair vs sales for comparable products. Since the sales ranges can differ a lot (from hundreds to millions) I would first have to scale them all to the same range, and preferably also shift them so the timeframes match. Not all products are sold at the same time, so at a certain moment I need to concert it from a pure calendar / timescale range to something like 'first period of sales, second period and so on ' so I can compare over a 2 year scale.

Hope I make some sense here :-)

I've been playing around already but get stuck in the complexity of it. So what would the best approach to tackle this? Not looking for ready-to-use code nut more of a advice towards common sense steps to take.