Kevin and I have been trying really hard to figure out which advertising medium to use, but it’s been kind of a tough problem figuring out which advertiser offers the best bang-for-our-buck.  I requested demographic numbers for both big radio companies in town (Reier and Gap West) as well as numbers for Bresnan cable advertising.  All data used in this study was for adults aged 35 and over.  Unfortunately the numbers came in slightly different formats so I had to jiggle things a little bit but in the end I think I got some reasonably decent comparisons.  The numbers we received didn’t tell us anything about probability to buy from each medium, though so I figured I would try to find out.

I devised a simple survey composed of five questions:

·         Would you be more likely to make a purchasing decision based on a radio advertisement, a television advertisement or neither? 

·         On average how many hours of television do you watch at work, per day?

·          On Average how many hours of radio do you listen to at work, per day? 

·         On average how many hours of radio do you listen to outside of work per day? 

·         And on average how many hours of television do you watch outside of work each day? 

I then created a random sampling of approximately 300 Gallatin Valley businesses, taken straight from the phone book.  I did have some problems with repeat numbers within my sample, or several different numbers that went to the same doctor’s office so in the end I ended up attempting to contact 260 unique businesses.  On each of these calls I told the person that answered the phone that I was conducting a survey and I asked if there was someone there who can make purchasing decisions, for the business, that would be willing to answer a couple of really quick questions?  Of the 260 businesses I called 46 agreed to take the survey.

Results

In case you don’t want to read through all the statistics I will present my findings right here.  First, people are dirty liars.   The average response to the question “On average how many hours of television do you watch outside of work each day?” was 1.426 conflicting pretty extremely with other, more reputable surveys found here and here suggesting the true average is closer to 2.5-4.5 hours.  Otherwise I didn’t find out anything else that was too ground-breaking.  The only two statistically significant comparisons I got were that people do not make purchasing decisions based on radio or television advertisements (Neither) more often than they make purchasing decisions based on a television advertisement.  And, on average, people listen to more hours of radio at work than they watch television at work.

How the data helped me

Using the confidence interval I obtained from the t-test I ran on “likeliness to purchase based on radio advertisement versus likeliness to purchase based on a television advertisement” I created a 95% high and low range for television advertising prices.  With the formula (1+(-.0762676))*television_advertising_cost I set the lower bound where 95% of the time the actual probability-to-purchase, weighted, cost per impression would be higher than this figure.  The upper bound was set with (1+.2936589)*television_advertising_cost representing a level where only 5% of the time the actual advertising cost would be higher.  I would have also liked to incorporate the fact that more radio is listened to during the work day than television is watched, however the demographic numbers we received from the advertising entities already had these figures at least partially figured in to their data so it was difficult to make a fair weighting system that didn’t compound the effects that were already seen.

How does the advertising stack up?

I did my best to include all of the current promotions for each company (Bresnan, Gap West and Reier) to ensure that everyone was on a an equal footing.  Because it is the summer, however, and fewer people are watching tv (I assume) Bresnan has been having some pretty crazy deals.  Their current special is for 500 all-day 30 second ads for $250.  Fifty cents per 30 second spot pretty much dominated the rankings and pushed cost per impression numbers below a cent for a couple of channels.  Omitting the specials and 95% range adjustments cost per impression numbers ranged from 2.4 cents for all day Fox News to 64 cents for KBOZ FM from 6am – 10 am.  Television and radio seemed fairly competitive, out of the top 10 non-special, non-adjusted best deals 4 were radio stations and 6 were television stations.

The Top 10 Deals (non-adjusted, non-special)

Download the full list here.  Without adjustments it looks like television, on average, is a slightly better deal than radio.  Adding in the probability-to-buy adjustments and the numbers get a lot closer with radio taking a small lead.  If I choose to pursue this further I think I will definitely want to find out different retention rates for radio and television advertisements. The rest of my data and R code can be downloaded from the links below.

Survey Data :::: Television Demographic Data from Bresnan :::: Radio Demographic Data from Gap West for Gap West and Reier :::: R Code.