| bigjon said: ok, I'm back. you do realize it is very possible to get within 5% with a 1% sample if you take they sample from unbiased representive sources. I will give you an example- When I was in college for a statistics project I got the data of types of cars people owned on my campus(ford, honda, etc) and my sample was VERY small compared to the total population(like .0000001%). But do you what I found? My sample was with 10% of that of the national statistics based hard retailer sales. 10% is huge is VGC really had that much. And yes NDP is probably perfect for their 60%, but the problem is that Walmart makes up for the rest. And since they are making an estimate to what walmart is selling but have no data they are making and nothing more than an educated guess. They are making this "guess" based on historical data, which I pointed out would be irrelevant in the changing marketplace. I believe walmart Walmart is throwing off their numbers by ALOT. I am not gonna pull numbers out of my rear, but I am gonna guess where you see NDP high and VGC low(and Vice Versa) you can bet that it would be somewhere in the middle. |
lets see...
1. yes, you can get an accurate representation of an accurate market with a sample of 1% and only have a 5% degree of error.
2. historical data is a good place to start when you dont have anything else to base your projections/ estimates or forecasts on. it isnt 100% accurate but as the saying goes "history always repeats itself". companies still look at and use historical data to project sales growth, profit growth, overhead expensses, etc etc.
3. in the end, all these estimates are all guesses but "intelligent guesses". none are 100% accurate thats for sure. and it isnt unheard of for some estimates to be off even with a larger sample size. its all guess work and during december, vgc just made a better estimate. january figures however had vgc adjusting figures to aline with npd data, probably coz vgc saw that the npd numbers were closer.
