Download MEDIA FLIGHT PLAN 7th Edition and more Exercises Printing in PDF only on Docsity!
MEDIA FLIGHT PLAN 7
th
Edition
1 st^ Printing Errata
CHECK YOUR TITLE PAGE! IF YOU HAVE Printing 2.0, you do NOT need this document, these
changes have already been made.
This document highlights several errors in the text that are critical to understanding the text or
to working exercises.
You can make the changes in the text yourself, however, we recommend that you either:
1) Print this document one-sided and then cut out the affected formulas and tables and
paste them in the appropriate place in your book. We recommend this option as this
will permanently correct your book without worry over lost printouts or typos. If you
choose this option, make sure you change the print setting in Adobe Acrobat to print
‘Actual size’ instead of ‘Fit’ or ‘Shrink oversized pages’ so that the replacement text will
cover the original.
2) Print this entire document two-sided (you will then have four sheets including this one)
and place the new sheets in your book.
The specific corrections are listed here— note that a few corrections below do not have pages
to reprint—you will need to write those in your book (See next page):
Pages 53/54:
Formula E#5] is incorrect.
In the table at the bottom of page 53, continuing to page 54, the combination notation is
inverted (the calculations are correct, however).
(over)
Page 62:
The formula following the sentence “ First, figure out the combination part,
:”^ is incorrect.
The ‘r’ in the denominator should be a ‘k’.
In the formula for Excel in the middle of the page, all the ‘r’s should be ‘k’s.
In the table at the bottom of the page, there should be a minus sign in the n!/k!(nk)! Heading
(e.g. it should read n!/k!(n-k)!)
Page 149:
The table for problem 7, the figures in the “Actions” column is incorrect.
THE FOLLOWING CHANGES YOU MUST WRITE IN YOUR BOOK, THERE ARE NO PAGES TO
PRINT:
Page 126:
Second line should read “in cell B4 ” not B5.
Page 128:
Last line of Step 24 should read “17.9 to 20.4”, not 17.
Page 129:
-Step 25, second bullet point should read (A30 to G40) , not (A45 to G55).
-Step 25, fourth bullet should read A45 not A
-Step 26, second bullet point should read same as Step 25, (A30 to G40) not (A45 to G55).
-Step 26, fourth bullet point should read A60 (different than step 25), not A73.
-Step 30, first bullet should read “Click File on the menu at the top of the screen…”, not click the
round “Windows Button”.
Page 155:
In part e) of problem 3, the heading in the table that ends (Use E#14) should read (Use E#16) as
it does in the problem text.
Page 157:
The last sentence of part a) for problem 4 is cut off. It should read:
(Use E#20 in Excel for the beta function).
Replacement pages start on next page.
Chapter 9
Issues Combinatorial Formula Result: Number of Ways
3 of 4
(^4) 4! 4 * 3* 2 *1 24
(^3) 3!(4 3)! 3* 2 1 (1) 6
=^ =^ = −
4 ways to see exactly 3 issues
4 of 4
=^ =^ =
1 way to see exactly 4 issues (recall(4-4)!= 0! = 1)
Now that all the ways one could be exposed to 0,1,2,3 or 4 issues are calculated, the probability of each level of exposure can be estimated using probability rules. Continuing with People magazine, with an average is- sue rating of 15.0:
Probability of reading Formula
Result (rounded) 0 of 4 Ex: not see #1 and not see#2 and not see #3 and not see # =(1.0-0.15) * (1.0-0.15)* (1.0-0.15)* (1.0-0.15) =0.850.850.85*0.85 or alternatively 0.15^0 *0.85 4
0.5220 or 52.20%
1 of 4 Ex: see #1 and not see#2 and not see #3 and not see # =0.15 * (1.0-0.15)* (1.0-0.15)* (1.0-0.15) =0.15 0.850.850.85 or alternatively 0.15^1 *** 0.85 3*
0.0921 or 9.21%
2 of 4 Ex: see #1 and see#2 and not see #3 and not see # =0.15 * 0.15 * (1.0-0.15)* (1.0-0.15) =0.15 0.150.850.85 or alternatively 0.15^2 *** 0.85 2*
0.0163 or 1.63%
3 of 4 Ex: see #1 and see#2 and see #3 and not see # =0.15 0 .15 0 .15 * (1.0-0.15) =0.15 0.150.150.85 or alternatively 0.15^3 *** 0.85 1*
0.0029 or 0.29%
4 of 4 Ex: see #1 and see#2 and see #3 and see # =0.15 * 0.15 0.15 * 0. =0.15 0.150.150.15 or alternatively 0.15^4 *** 0.85**^0
0.0005or 0.05%
Notice the pattern in the alternative formulas: probability of seeing# to see^ x probability of not seeing # not to see^.
If we let ‘p’ represent the probability of seeing, this can be written:
E#6] ( 1 ) k n^ k
p p
−
This looks similar to the last part of the BBD formula presented in the introduction. It was partly this similar- ity which led to combining the beta function with the binomial expansion.
The binomial expansion (multiplying the ‘number of ways’ with the corresponding probability) is used to cre- ate the complete frequency distribution. Let’s multiply them together and see if it all works. A correct result will sum to 100% (e.g. summing all the frequency cells should equal 100%):
Chapter 9
β (a, b+n ) from Excel = exp(gammaln(a)+gammaln( b+n )-gammaln(a+ b+n )) =exp(gammaln(0.161)+gammaln( 1.200+1 )-gammln(0.161+ 1.200+1 )) = 5.
Important! Notice that in the numerator, ‘b+n’ is used for ‘b’ every time! If you do not make this substitution where required in the beta function you will get an incorrect result.
Denominator:
β (a,b) from Excel = exp(gammaln(a)+gammaln(b)-gammaln(a+b)) = exp( gammaln(0.161)+ gammaln(1.200)-gammaln(0.161+1.200)) = 5.
( ) ( )
C1 1 1 1 0.882 0.118 or1 1.8%
a b
a b
β β
The cumulative reach for each of the remaining successive issues is calculated the same way, substituting the correct number of issues for ‘n’ each time. Here are all 6 issue results compared to the measured values:
Cume Reach
Issues
Politz
Study
BBD
Model Probability
Notice the BBD model reach estimates are quite close to the tabulated values. Commercial programs used in agencies make minor tweaks to further improve the fit. The basic probability estimates are also shown in the table and graph—and rapidly stray from the measured estimates. BBD does much better because it uses more information, and does not assume independence (e.g. BBD can accommodate duplication that is not random, whereas the assumption of random duplication is explicit in basic probability—when removing the duplication back in E#3 , the (p 1 * p 2 ) estimate of duplication assumes randomness because the two events are assumed independent).
Let’s try calculating the frequency distribution for six issues. The sum of the exposure cells 1 through 6 must total the 28.7% (cumulative reach) estimated for the six issues, so that figure can be used to verify the results.
E#13 can be used to calculate the entire frequency distribution:
Read k of n issues=
( ) ( )
n a k b n k
k a b
β β
+^ +^ −
Exercise 18
- You’ve received the following sales data by website from your provider. Complete the CPA analysis:
Website
Delivered Impressions Actions
Cost of Impressions
CPA
(Cost-Per-Action) ESPN.com 2,000,000 3,750 $30, Eonline 1,000,000 1,100 $15, Zillow.com 1,500,000 800 $ 7, AllRecipies.com 500,000 500 $ 3, History.com 1,200,000 1,400 $ 8, iVillage.com 750,000 200 $ 3,
- Which three websites will you recommend to receive the extra funds, and why?
9.In the table below, indicate with a check which two websites you would eliminate based on each of the CTR, CPC and CPA criteria. Explain why you might want to consider different criteria, instead of say, focusing just on CTR?
Website CTR CPC CPA ESPN.com Eonline Zillow.com AllRecipies.com History.com iVillage.com
Exercise 18
- In digital as in other media, we use indices to indicate the likelihood that our target exhibits specific be- haviors online or offline (e.g., watch videos online or visit Pandora). In the Digital Media chapter, you learned about defining targets. Running reports and interpreting them is an important part of the process. The Delorme client has decided to target 18-34 year olds as they are especially active in outdoor activities and make good prospects for their handheld GPS devices.
Your data provider has issued report, but parts were illegible when they came through on the fax machine. You realize that there is sufficient data on the report to fill in the missing data, and decide to do it by hand rather than loose time by having them re-run it. Complete the report by calculating the missing data and indexes. (The target audience for this report is adults 18-34):
Behavior
Total Base Audience
Percent Vertical of Total
Target Audience
Percent Vertical of Target Index BASE- 18+ Years Old 181,628,000 100.0 25,416,000 100.0 100 Used Social Media Past 30 Days 139,480,000 76.8 21,650,000 85.2 111 Watch 1-5 video streams on average day 69,669,000 38.4 12,254,000 48. Visit Newspaper site Past 30 Days 58,243,000 32.1 6,244,000 24. Played Gaming Console Past 30 Days 42,890,000 23.6 10,203, Streamed Pandora Past 30 Days 15,743,000 4,751,
- Based on your calculations, which kinds of sites (social media like twitter, facebook; video streaming like YouTube,etc; news sites; gaming sites or streaming audio like Pandora) would you be most likely to recom- mend?
- Which kind of site does not perform well against the 18-34 year old target?
- What does the index for “Streamed Pandora Past 30 days” mean? The index for newspaper sites? Be specific—which direction from average and by how much. (Hint: read the section Calculating Indexes in the chapter Using Indexes to Identify Target Audiences if you need help)
