
Advanced Statistics Advanced Probability and Statistics Math Forum 
 LinkBack  Thread Tools  Display Modes 
September 12th, 2012, 08:44 AM  #1 
Member Joined: Aug 2012 Posts: 41 Thanks: 0  Sales forecasting
Ok... I'm trying to brush off my (arguably bad) math skills before going back to uni by taking on problems that directly affects me. One of the challenge I have as an owner of a sales organization is to test script and new offers constantly. What I do is usually test a sample (ideally, I would make a lot of pitches to see if the offer works), before giving it out to sales reps. My question:How do I determine how large a sample needs to be to be before knowing I can replicate the results confidently? The smaller the better obviously. I would like to get to the point where I can confidently say: "For this amount of people pitched, for this demographic ( say small business owners), I will get this this this amount of sales for the next years to come" I'm understand it's rather vague...I'm not looking nescessarily for complete answers... just pointers so I can look in the right direction. I wikied "sales forecasting" but it seems heavily theoryoriented,does any of this works in practice? 
September 13th, 2012, 04:55 AM  #2 
Senior Member Joined: Aug 2012 Posts: 229 Thanks: 3  Re: Sales forecasting
Hey spook. So in terms of forecasting, what we typically look at is a model and a common kind of model is known as a regression model. But you will need to describe what kinds of data you have (as best you can publically of course), because depending on the kind of data and its own model, it will affect how you do things. Now regardless of that, here are some really important issues to consider: 1) How far into the future you are trying to make forecasts for (with respect to the start and end points of your data) 2) The nature of the data for the forecast (sudden changes in situations would make forecasts absolutely useless and if these happen frequently, you need to be honest in considering them) 3) Other things like economic fundamentals, market fundamentals, and so on. A big example of three is that continual growth is just not feasible: it looks nice on an excel chart but it's just nuts. Even if you do keep growing some time, all it takes for you to be catapulted down is a slowdown, a crash, a popped bubble, or some really major event tied to the thing you are selling (like say the stuff you sell is made with material X or in place Y and suddenly the prices shoot up, less people buy and if you rely on growth, you'll be completely unprepared). Now if you are talking about years, then I think you should maybe play it a bit safer and consider a more short term model. The economy is really volatile at the moment and things are crazy right now world wide. So in terms of the simplest kind of regression model, you basically use a program like a statistical program or something like in Excel to generate a regression and then you can generate a fit given your existing data which can help you move towards forecasting: it's not just run a program generate a fit and use that, it's an iterative process that takes into account what you know as an expert, what your data is showing, and how you end up combining all that to get something thats reasonable. So yeah the most basic framework is a multiple linear regression, and you can transform your data and create whatever kinds of terms you want (like x^2 or SQRT(x)) to fit this kind of model with the same framework and commands. Finance, as you may imagine, devotes a lot of time looking at forecasting (for obvious reasons), but my advice is to just keep it simple and that is basically seeing what a linear regression is first. Good luck! 

Tags 
forecasting, sales 
Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
Forecasting percentages by proportions  haz33  Elementary Math  3  March 9th, 2014 09:57 PM 
Sales conversion linked to sales enquiries and orders  Rocky  Elementary Math  1  July 3rd, 2013 08:55 AM 
Not sure about the increased sales target  tomb  Economics  5  November 16th, 2011 10:06 AM 
Customer sales  Dld3  Economics  3  November 2nd, 2011 11:17 AM 
Difficult forecasting/ density estimation problem  whaddup  Advanced Statistics  1  July 4th, 2011 04:59 AM 