Every Appstate business graduate could figure this one out:
A consultant to the state lottery said in a recent report that spending on ads is “highly related” to sales. Recent figures in North Carolina don’t bear that out. For example, ad spending jumped by 17 percent from the 2009-10 fiscal year to the next. But sales increased only 2.8 percent. The next year, ad sales increased by a more modest 6.5 percent, but gross sales jumped by nearly 9 percent.
Here are the most recent numbers at a glance:
Fiscal year Ad spending Ad spending change Sales Sales change 2007-08 $8.6 million – $1.07 billion – 2008-09 $11.2 million 30.2% $1.29 billion 20.6% 2009-10 $11.8 million 5.4% $1.42 billion 10% 2010-11 $13.8 million 16.9% $1.46 billion 2.8% 2011-12 $14.7 million 6.5% $1.59 billion 8.9% 2012-13 $15.3 million 4.1% $1.69 billion 6.3%
Source: N.C. Education Lottery, Delehanty Consulting LLC
The regression of sales on ad spending has a R-squared of 0.96 (n=6, my guess is that the consultant used more data than this):
Given this model, a doubling of ad spending (from $15.3 million to $30.6 million) would lead to an increase of $1.31 billion in sales. About 28% of this is $385 million. This is about twice enough to pay for a 5% increase in teacher salaries (95,000 teachers earn about $41,000 annually [pdf]).
The problem, however, is that this is a way out-of-sample forecast. It is based on a huge increase in advertising. Someone should keep that in mind. And, even with the larger sample probably used by the consultant, it is still very small with a lot of other variables floating around that should be considered.