Licensee Badge
Submitted by intp89 on


I am trying to develop an MT goal related to a model we are tasked to build that will predict our market share in the upcoming year.  T is easy.  I know when it needs to be ready.  I think an obvious measure would be, predict next year's market share within +/- X%.  I can work with my team and boss about what the X should be.  

But here is the rub.  Our prediction will rely on future business implementations by our company, which is often hard to quantify.  Real life example here.  Say we forecast the market share, the company plans around that, and then a decision is made to greatly reduce the marketing budget.  The fact that we "do marketing" is a variable in the model, but it is a little more art than science.  How would I handle at end of year?  Should I say "Go back and run the model and make an adjustment to reflect our reduced focus on marketing.  Then calculate how accurate we would have been"?  Since the way the "impact of marketing" is incorporated into our model in the first place relies on judgement calls (every marketing variable cannot be quantified or the model would have so many inputs as to make it useless), it will be hard not to be biased on that adjustment.  

Am I overthinking this? After re-reading my post, I'm guessing that I am?  Maybe I can just address it when doing the annual review of my analyst charged with this as his primary responsiblity, and hope my boss will do the same when he evaluates me?