Академический Документы
Профессиональный Документы
Культура Документы
1
Zip Code Avg. Price Quantity Begin with
48104 18.78 55 only a small
43231 18.99 54 amount of
60202 20.87 23 information ...
70803 20.23 38 sales and
90048 18.99 28 price data
30064 18.46 47 broken down
95616 19.58 35
by zip code.
32611 20.91 37
94720 19.45 33
50311 19.77 35 The manufacturer
80231 20.45 30 has these data if
it tracks where its
20902 20.02 41 customers live,
93106 22.45 2 and what they paid.
45244 18.55 51
(Here average price is measured in thousands of $.)
2
A Few Simple Statistics ...
Average sales per zip code = 32.84.
Median sales = 33.
Average price per mini-van = 20.11 (or $20,110).
Median price = 20.04.
3
ANNUAL DEMAND FOR THE FIRM'S MINI-VANS
25
20
PRICE ($1000s)
15
Series1
10
0
0 10 20 30 40 50 60
Chrysler’s price: The average price (in $1000s) of the Chrysler mini-vans
sold in this zip code.
5
We can easily get more data ...
Location Zip Avg. Price Quantity Incom e Population % Under 18 Chrysler "Advertising"
Ann Arbor 48104 18.78 55 36.67 20.01 25.67 22.97 0.41
Westerville 43231 18.99 54 29.33 24.46 33.24 19.17 4.83
Evanston 60202 20.87 23 36.97 17.67 23.13 19.45 3.94
Baton Rouge 70803 20.23 38 30.97 17.94 29.33 20.36 2.56
Los Angeles 90048 18.99 28 32.65 25.88 17.97 18.76 0.21
Marietta 30064 18.46 47 27.85 20.77 30.76 20.01 3.67
Davis 95616 19.58 35 36.52 21.32 22.67 20.74 0.32
Gainesville 32611 20.91 37 34.24 21.76 28.36 19.88 0.53
Berkeley 94720 19.45 33 34.64 23.58 29.47 18.66 0.35
Des Moines 50311 19.77 35 27.36 21.46 28.93 18.87 5.23
Denver 80231 20.45 30 27.87 21.53 28.39 19.19 3.82
Wheaton 20902 20.02 41 28.76 22.87 29.82 19.01 4.91
Santa Barbara 93106 22.45 2 43.66 11.62 15.12 18.25 0.88
Cincinnati 45244 18.55 51 29.02 22.51 31.85 19.77 3.75
6
Location Zip Avg. Price Quantity Incom e Population % Under 18 Chrysler "Advertising"
Ann Arbor 48104 18.78 55 36.67 20.01 25.67 22.97 0.41
Westerville 43231 18.99 54 29.33 24.46 33.24 19.17 4.83
Evanston 60202 20.87 23 36.97 17.67 23.13 19.45 3.94
Baton Rouge 70803 20.23 38 30.97 17.94 29.33 20.36 2.56
Los Angeles 90048 18.99 28 32.65 25.88 17.97 18.76 0.21
Marietta 30064 18.46 47 27.85 20.77 30.76 20.01 3.67
Davis 95616 19.58 35 36.52 21.32 22.67 20.74 0.32
Gainesville 32611 20.91 37 34.24 21.76 28.36 19.88 0.53
Berkeley 94720 19.45 33 34.64 23.58 29.47 18.66 0.35
Des Moines 50311 19.77 35 27.36 21.46 28.93 18.87 5.23
Denver 80231 20.45 30 27.87 21.53 28.39 19.19 3.82
Wheaton 20902 20.02 41 28.76 22.87 29.82 19.01 4.91
Santa Barbara 93106 22.45 2 43.66 11.62 15.12 18.25 0.88
Cincinnati 45244 18.55 51 29.02 22.51 31.85 19.77 3.75
45
40
IN COME ($1000s)
35
30
25 Series1
20
15
10
5
0
0 10 20 30 40 50 60
SALES
30
25
20
Series1
15
10
0
0 10 20 30 40 50 60
SALES
35
30
25
20
Series1
15
10
0
0 10 20 30 40 50 60
SALES
5
ADVERTISING ($1000s)
Series1
3
0
0 10 20 30 40 50 60
SALES
25
CHRYSLER'S PRICE ($1000s)
20
15
Series1
10
0
0 10 20 30 40 50 60
SALES
13
Our objective:
Isolate the effect price has on sales,
holding all other variables constant.
But how?
14
What happens if I divide my zip
codes up so that they have very
similar residents?
Location Category # Zip Avg. Price Q Income Population % < 18 Ad$ Chrysler
Evanston, IL Affluent 3 60202 20.87 23 36.97 17.67 23.13 3.94 19.45
Santa Barbara, CA Affluent 13 93106 22.45 2 43.66 11.62 15.12 0.88 18.25
Water Mill, NY Affluent 15 11976 23.01 7 41.35 10.11 19.36 3.44 20.23
Irvine, CA Affluent 17 92717 19.95 20 39.61 14.37 24.12 0.87 18.45
Park Ridge, IL Affluent 18 60068 20.76 24 37.89 14.39 28.45 4.62 19.18
San Francisco, CA Affluent 20 94102 20.34 12 39.88 16.63 17.52 1.32 18.15
Rutherford, NJ Affluent 26 7070 20.06 25 37.38 17.36 20.46 4.37 20.77
New Rochelle, NY Affluent 43 10801 21.22 21 40.85 16.93 19.47 3.88 19.48
15
We can create other
categories, too….
• College towns
• Urban/mixed
• Zip codes with many young families
• “Other”
16
What happens if I divide my zip
codes up so that they have very
similar residents?
Category Income Price Sales Population % < 18 Chrysler Adve
Affluent 39.7 21.1 16.8 14.9 21 19.2 2
College Town 34.4 20.4 35.9 18.8 26 20.9 1
Urban/Mixed 30.6 20.5 28.1 22.6 27 19.1 0
Families 28.4 19 43.7 22.2 31 19.5 4
Notice that affluent zip codes share more than just high income.
They also tend to have lower populations and a smaller fraction of children.
This does not completely solve the problem, but it helps.
17
INCOME AND SALES -- AFFLUENT ZIP CODES
44 Here is the
42
cause of the
IN COME
40 Series1
38
negative
36 relationship!
0 5 10 15 20 25
SALES
40
IN COME
30
20 Series1
10
0
0 10 20 30 40 50 60
SALES
18
INCOME AND SALES -- URBAN/MIXED ZIP CODES
40
30
IN COME
20 Series1
10
0
0 5 10 15 20 25 30 35
SALES
32
31
30
Income
29 Series1
28
27
26
0 10 20 30 40 50 60
Sales
19
Must we limit ourselves to
looking at just two
variables at a time, or is
there some way to
express a relationship
between all of the
variables?
20
Expressing an “over-arching”
mathematical relationship:
Estimated Quantity = - 32.5
- 4.22*(Price)
+ 0.31*(Income)
+ 0.92*(Population)
+ 0.94*(Advertising)
+ 4.52*(Chrysler’s Price)
21
Application #1:
forecasting demand in new markets
Estimated Quantity = - 32.5
- 4.23*(Price)
Suppose the manufacturer
enters a new region of + 0.32*(Income) 34.40
the country. Among the
new zip codes where it
+ 0.92*(Population) 18.8
expects sales is one with 26.0
the characteristics here.
+ 1.13*(% of Population < 18 Years)
+ 0.94*(Advertising) 1.10
22
Application #1:
forecasting demand in new markets
Estimated Quantity = - 32.5
- 4.23*(Price)
If we substitute all of this
information into the + 0.32*(Income) 34.40
demand relationship, we
can express sales as a
+ 0.92*(Population) 18.8
function of price alone ... 26.0
+ 1.13*(% of Population < 18 Years)
+ 0.94*(Advertising) 1.10
28.8
Sales in
121.52 this zip code
24
Application #2: forecasting
demand in existing markets
Estimated Quantity = - 32.5
- 4.23*(Price)
Suppose the manufacturer
wants to know the relation- + 0.32*(Income) 29.33
ship between price and
sales in Westerville, Ohio,
+ 0.92*(Population) 24.46
a market it already serves. 33.24
Then ...
+ 1.13*(% of Population < 18 Years)
+ 0.94*(Advertising) 4.83
- 4.23*(Price)
But this reflects demand in +2.0
Westerville, Ohio LAST + 0.32*(Income) 29.33
YEAR. Suppose that this
year Westerville looks the
+ 0.92*(Population) 24.46
same, except that it has 33.24
2000 more people. Then ...
+ 1.13*(% of Population < 18 Years)
+ 0.94*(Advertising) 4.83
Q sold in
Q1 Q2 128.7 130 this zip code
27
Application #3: Assessing the
impact of a change in a rival’s price
Estimated Quantity = - 32.5
- 4.23*(Price)
Now ... suppose that in +2.0
addition to the 2000 new + 0.32*(Income) 29.33
people, Chrysler lowers its
price by $1800 on average.
+ 0.92*(Population) 24.46
Then ... 33.24
+ 1.13*(% of Population < 18 Years)
Q Sold in
121.8 this zip code
29
Application #4: assessing the
performance of individual dealers
Estimated Quantity = - 32.5
- 4.23*(Price)
The key word here
is “estimated.” We + 0.32*(Income)
can not forecast
sales perfectly. + 0.92*(Population)
+ 0.94*(Advertising)
+ 4.52*(Chrysler’s Price)
30
Let’s look at the table again ...
Location Income Price Population % < 18 Chrysler Ad $ Quantity Estimate Difference
Evanston, IL 36.97 20.87 17.67 23.13 19.45 3.94 23 25 -2
Santa Barbara 43.66 22.45 11.62 15.12 18.25 0.88 2 -3 5
Water Mill, NY 41.35 23.01 10.11 19.36 20.23 3.44 7 9 -2
Irvine, CA 39.61 19.95 14.37 24.12 18.45 0.87 20 20 0
Ann Arbor, MI 36.67 18.78 20.01 25.67 22.97 0.41 55 51 4
BatonRouge, LA 30.97 20.23 17.94 29.33 20.36 2.56 38 36 2
Ithaca, NY 37.83 22.01 16.94 21.45 21.23 0.26 28 22 6
Philadelphia, PA 27.45 21.57 22.07 30.11 21.69 0.63 25 38 -13
Boston, MA 27.45 22.03 21.87 24.64 19.01 0.57 14 18 -4
Atlanta, GA 31.17 21.04 23.18 30.11 18.85 0.93 35 29 6
Westerville, OH 29.33 18.99 24.46 33.24 19.17 4.83 54 47 7
Marietta, GA 27.85 18.46 20.77 30.76 18.85 3.67 47 41 6
Des Moines, IA 27.36 19.77 21.46 28.93 20.43 5.23 35 42 -7
Cincinnati, OH 29.02 18.55 22.51 31.85 18.87 3.75 51 43 8
Nashville, TN 33.57 20.01 24.22 27.56 20.21 0.25 26 38 -12
Holland, MI 29.47 18.95 24.16 29.34 18.78 0.67 45 37 8
31
The first of these columns takes
information from each row ...
Location Income Price Population % < 18 Chrysler Ad $ Quantity Estimate Difference
Evanston, IL 36.97 20.87 17.67 23.13 19.45 3.94 23 25 -2
Santa Barbara 43.66 22.45 11.62 15.12 18.25 0.88 2 -3 5
Water Mill, NY 41.35 23.01 10.11 19.36 20.23 3.44 7 9 -2
Irvine, CA 39.61 19.95 14.37 24.12 18.45 0.87 20 20 0
Ann Arbor, MI 36.67 18.78 20.01 25.67 22.97 0.41 55 51 4
BatonRouge, LA 30.97 20.23 17.94 29.33 20.36 2.56 38 36 2
Ithaca, NY 37.83 22.01 16.94 21.45 21.23 0.26 28 22 6
Philadelphia, PA 27.45 21.57 22.07 30.11 21.69 0.63 25 38 -13
Boston, MA 27.45 22.03 21.87 24.64 19.01 0.57 14 18 -4
Atlanta, GA 31.17 21.04 23.18 30.11 18.85 0.93 35 29 6
Westerville, OH 29.33 18.99 24.46 33.24 19.17 4.83 54 47 7
Marietta, GA 27.85 18.46 20.77 30.76 18.85 3.67 47 41 6
Des Moines, IA 27.36 19.77 21.46 28.93 20.43 5.23 35 42 -7
Cincinnati, OH 29.02 18.55 22.51 31.85 18.87 3.75 51 43 8
Nashville, TN 33.57 20.01 24.22 27.56 20.21 0.25 26 38 -12
Holland, MI 29.47 18.95 24.16 29.34 18.78 0.67 45 37 8
33
We can then go to our dealers in
Philadelphia and Nashville ...
Location Income Price Population % < 18 Chrysler Ad $ Quantity Estimate Difference
Evanston, IL 36.97 20.87 17.67 23.13 19.45 3.94 23 25 -2
Santa Barbara 43.66 22.45 11.62 15.12 18.25 0.88 2 -3 5
Water Mill, NY 41.35 23.01 10.11 19.36 20.23 3.44 7 9 -2
Irvine, CA 39.61 19.95 14.37 24.12 18.45 0.87 20 20 0
Ann Arbor, MI 36.67 18.78 20.01 25.67 22.97 0.41 55 51 4
BatonRouge, LA 30.97 20.23 17.94 29.33 20.36 2.56 38 36 2
Ithaca, NY 37.83 22.01 16.94 21.45 21.23 0.26 28 22 6
Philadelphia, PA 27.45 21.57 22.07 30.11 21.69 0.63 25 38 -13
Boston, MA 27.45 22.03 21.87 24.64 19.01 0.57 14 18 -4
Atlanta, GA 31.17 21.04 23.18 30.11 18.85 0.93 35 29 6
Westerville, OH 29.33 18.99 24.46 33.24 19.17 4.83 54 47 7
Marietta, GA 27.85 18.46 20.77 30.76 18.85 3.67 47 41 6
Des Moines, IA 27.36 19.77 21.46 28.93 20.43 5.23 35 42 -7
Cincinnati, OH 29.02 18.55 22.51 31.85 18.87 3.75 51 43 8
Nashville, TN 33.57 20.01 24.22 27.56 20.21 0.25 26 38 -12
Holland, MI 29.47 18.95 24.16 29.34 18.78 0.67 45 37 8
34
Let’s revisit the relationship between sales
and advertising expenditures ...
THE RELATIONSHIP BETWEEN SALES AND ADVERTISING
5
ADVERTISING ($1000s)
Series1
3
0
0 10 20 30 40 50 60
SALES
- 4.23*(Price)
+ 0.32*(Income)
+ 0.92*(Population)
+ 0.94*(Advertising)
+ 4.52*(Chrysler’s Price)