Marketing Optimization
Modeling
Theme Park Case Study
Overview and Disclaimer
• The following is a real case study from a
regional theme park. Data, labels and
brand names have been masked to
preserve confidentiality.
Situation
The client is a small regional theme park. It’s core marketing budget
is modest, about $2.5M and is spread across outdoor, radio and print
advertising. Apart from this, there is a special TV advertising effort in
November and December to promote the park’s special Holiday Lights
program. In the past year, overall attendance has been flat. Theme
parks are very seasonal businesses and are significantly impacted by
short-term weather patterns. For the past year, the management of
the theme park has been concerned about the impact of rising fuel
prices. Their main interest is to be able to understand how to
leverage their modest marketing budget for re-igniting attendance
growth at the park.
Outline
1. Model Architecture
2. Decomposition of Attendance Drivers
3. Marketing Variance Analysis
4. Seasonal and Days-of-Week Effects
5. Sensitivity Analyses
1. Media
2. Gate Price
3. Fuel Prices
4. Weather
6. Impact of Promotional Events
7. Marketing Efficiencies: Revenue per Expenditure
8. Marketing Spending Optimization
9. Media Schedule Optimization
10. Attendance Simulation and Forecast
11. Model Validation
Model Architecture
Media Outdoor, Print, Radio, Magazines
Weather Temperature Maximum Daily Temperature
Atten- Apr04-Dec06
dance By Day Weather Sunshine, Precipitation
Conditions
Promotional Events Special “Themed” Events and Programs
Auto Fuel Prices Weekly average pump price
Daily park attendance as
affected by media, Seasonality Month and Day of Week
promotional events,
Gate Price Attendance paid price
pricing, weather, holidays,
and fuel prices.
Holiday Lights TV Annual special holiday event TV adv., Nov
24-Jan 4
Decomposition of Theme Park Annual Attendance
For the past year, total media, including Holiday Lights, drove 13.8
percent of total attendance at the park. “Themed” Promotional
Events were responsible for driving 2.8 percent of total
attendance. The largest contributor was print media, followed by
Holiday Lights TV
Promo.Events
2.8%
Holiday Lights TV
4.8%
Base Other
OOH.Adv
83.4% 16.6%
2.5%
Radio.Adv
0.2%
Print.Adv
6.0%
Mag.Adv
0.3%
Decomposition of Theme Park Attendance by Week
140,000 Base Holiday Lights TV
Events & Promos OOH.Adv
Radio,.Adv Print.Adv
120,000 M ag.Adv
100,000
80,000
60,000
40,000
20,000
0
4/7/2004 7/6/2004 10/4/2004 1/2/2005 4/2/2005 7/1/2005 9/29/200 12/28/200 3/28/200 6/26/200 9/24/200
Marketing Variance Analysis: Drivers of Annual Attendance Trends Oct ‘06
Total park attendance for the year was basically flat at -0.6 percent.
Improved performance from Events and Promotions and Print media were
the largest positive contributors for the year. Cooler weather, a 25%
increase in fuel prices and a -91% reduction in radio advertising were the
key negative contributors year-over year.
6.9%
5.1%
2.8% Events and Promotions
2.4% Print Advertising
Base Momentum
1.6% Holiday Lights TV
1.1% Weather - Precipitation
0.3% OOH Advertising
-3.7% Magazine Adv
Attendance Pricing
-4.6%
Radio Advertising
-4.6% Gas.Prices
-7.9% Weather - Temperature
-10% -5% 0% 5% 10%
Monthly Seasonal Effects
Park attendance is highly seasonal, with peaks occurring in Spring
(April) and an uptick in the early Fall (October)
900
812
800
684 JAN FEB
700
600 MAR APR
500 483
451 MAY JUNE
422
400
400 347 JUL AUG
298 303
300
225 233
212 SEP OCT
200
100 NOV DEC
0
Theme Park “Day of the Week” Effect
There is a recurring pattern of attendance for days-of-the-week.
Attendance peaks on Thursdays due to regular 25% admissions
price discounts.
800
709
700 MON T UE
600 532 WED T HUR
473 FRI SAT
500 457
437
370 SUN
400 363
300
200
100
0
Attendance
Theme Park Media Sensitivities
On a dollar-for-dollar basis, radio and print are the most
responsive media forms, with magazine and outdoor being
considerably less responsive.
Daily Attendance
3,400
3,300
3,200
RADIO
PRINT
3,100
OUTDOOR
3,000 MAGAZINE
2,900
2,800
$0 $50 $100 $150 $200 $250
Monthly Spending $000
Gate Admissions Price and Average Daily Attendance
Raising the gate admissions price from $10.00 to $10.50 cost the park about -3.7% in daily
attendance
Daily Attendance
3,600
3,539
3,500
3,470
3,400 3,404
3,340
3,300 3,280
3,222
3,200
3,166
3,113
3,100
3,062
3,000 3,012
2,965
2,900
$9.00 $9.25 $9.50 $9.75 $10.00 $10.25 $10.50 $10.75 $11.00 $11.25 $11.50 $11.75 $12.00
Admissions Price
Impact of Auto Fuel Prices on Theme Park Daily Attendance
Rising auto fuel prices impact attendance at the park and pose a future risk to increasing
attendance. This effect is of slightly greater importance than that of gate prices.
Daily Attendance
3,500
3,450 3,440
3,400
3,350 3,359
3,300 3,287
3,250
3,222
3,200
3,162
3,150
3,100 3,108
3,059
3,050
3,013
3,000
2,950
$1.80 $2.30 $2.80 $3.30 $3.80
Average Pump Price for
Regular Self-Serve
Temperature & Weather Impact on Daily Attendance
Short-term weather conditions of temperature and precipitation have a profound impact on daily
attendance at the Theme Park.
Daily Attendance
12,000
10,000 SUNNY
PARTLY/MOSTLY
CLOUDY
8,000 RAIN
SOME SNOW
6,000
SIGNIFICANT SNOW
4,528
4,000 3,768
3,065
2,423
2,000 1,843
1,619
1,079 1,330
0
15 25 35 45 55 65 75 85 95
Maximum Daily
Temperature
Impact of Special Events and Holidays on Theme Park Daily Attendance
Special park theme events have played an increasingly important role in driving incremental
attendance at the park. Easter and Holiday Lights events were of particular importance. Over the
past year, average attendance gains from these events more than doubled.
Incremental Daily Attendance
African Adventures
7,000 Mom's Day
Bug.Mania
6,000 Butterfly.Blast
Cats 'n Dogs
Dads Day
5,000 Easter Eggstream
Explorers Weekend
4,000 Grandparents' Day
Hispanic Weekend
Independence Day
3,000 Jungle.Day
Kids.Day
2,000 Play.Week
Retro.Weekend
Women's Day
1,000 Holiday Lights
2006 Average
0 2005 Aveerage
Attendance Lift
Theme Park Media Revenue per Dollar of Ad Investment
Because our models were built from individual media vehicles used, we are able to isolate the
effects of each media vehicle. In the past year, 35 percent of the media vehicles generated greater
than $1 break-even revenue. This compares unfavorably to the past year, when 65% of the media
vehicles generated greater than $1 per dollar investment. Cutting radio advertising in the current
year eliminated one of the most efficient and effective media vehicles.
Break-Even
MAG.KIDSTIMES '05
P rt.HISP ANICNEWS '05
P rt.TIMES '05
P rt.TRIBUNE '05
P rt.DAILYP OST '05
P rt.P RESS '05
P rt.THE RESIDENT '05
P rt.P RINTCTP OST '05
P rt.THERECORD '05
P rt.TRIBUNE '05
P rt.JOURNALNEWS '05
P rt.TIMES '05
P rt.EXP RESS '05
P rt.DAILYNEWS '05
RADIO.METRO '05
OOH.HWY.BULLETINS '05
OOH.WALLBOARDS '05
OOH.BUSSHELTERS '05
OOH.BRANDP OSTERS '05
OOH.P OSTERS '05
MAG.KIDSTIMES '06
P rt.TIMES '06
P rt.THE RESIDENT '06
P rt.DAILYP OST '06
P rt.TRIBUNE '06
P rt.HISP ANICNEWS '06
P rt.EXP RESS '06
P rt.JOURNALNEWS '06
P rt.TIMES '06
P rt.P RESS '06
P rt.DAILYNEWS '06
RADIO.METRO '06
OOH.HWY.BULLETINS
OOH.BUSSHELTERS '06
P rt.DAILYRECORD '06
OOH.BRANDP OSTERS '06
OOH.BULLETINS '06
$0.00 $7.00 $14.00
Incremental Impact on Attendance from a $100 increase in Media Spending
Radio and print media stand out as generating the most incremental attendance per dollar
investment.
164
168
161 Radio
154 132 Adv
147
140
133
126
119 Print Adv
112
105
98
91
84
77 Mag Adv
70 55
63 44
56
49
42
35 OOH
28 Adv
21
14
7
0
Theme Park Current and Optimal Marketing Investments
When comparing the investment in the four media, we see that OOH and Magazines generate
proportionately less attendance than their share of spending, while Holiday Lights, Radio and Print
media generate more. Thus our optimized solution calls for more spending for Holiday Lights, print
and radio and less for the other media. At constant total spending, this shift to a more efficient
media mix is estimated to generate +13 percent higher total attendance.
100%
80%
60%
40%
20%
0%
Attendance Current Optimal
Contribution Investment Investment
RADIO PRINT OUT DOOR MAGAZINE HOLIDAY LIGHT S
Theme Park Total Attendance and Returns per Media Investment
Presently, the park’s marketing generated about $1.56 in revenue per dollar of marketing
investment. This above “break-even” productivity can still be maintained up to +20% of current
spending levels. Still, the dilemma is that this $1.56 impact was a -22% decline from the prior year
and strongly supports an effort to improve and optimize spending efficiencies going forward.
$3.00 2,150,000
Revenue per $1 media investment Annual Attendance
$2.60 2,100,000
$2.50
$2.24
2,050,000
$2.00 $1.99
$1.78
Prior 2,000,000
$1.63 Net Returns to Marketing
$1.50 Year $1.47
Total Attendance
$1.32 1,950,000
$1.17
$1.00 $1.03
Current $0.92
$0.82 1,900,000
Year
$0.50
1,850,000
$0.00 1,800,000
,0 00 ,0 00 ,0 00 ,0 00 ,0 00 ,0 00 ,0 00
00 00 00 00 00 00 00
$ 1,0 $ 1,5 $ 2,0 $ 2,5 $ 3,0 $ 3,5 $ 4,0
Optimizing the Media Schedule
• The next two charts compare current versus an optimized schedule. The optimized schedule
places each media, as per the overall mix optimization, in the weeks such that its overall
impact on attendance will be maximized.
• This schedule calls for increases in print, radio and Holiday Lights advertising, and reductions
in magazine and outdoor media.
• The overall effect of this schedule optimization is to place more “weeks-of-execution” for
media in a more continuous fashion. Doing so, per the plan outlined, is expected to increase
attendance +1.5% with the current and constant dollars.
Current Media Schedule
The current media plan does not begin until April, is heavy on outdoor advertising and
has about a 10 week hiatus in the Fall.
$100,000
$90,000
$80,000
$70,000
HOLIDAY.LIGHTS
$60,000 MAG
$50,000 PRINT
$40,000 RADIO
$30,000 OOH
$20,000
$10,000
$-
0
3
6
9
2
5
8
1
4
7
0
3
6
9
2
k1
k4
W 7
k
k1
k1
k1
k1
k2
k2
k2
k3
k3
k3
k4
k4
k4
k4
k5
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
ee
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
W
Optimized Media Schedule
The optimized plan calls for more spending on print, radio and Holiday Lights advertising.
Overall, compared to the current plan, each media is placed into more total weeks and
advertising begins three weeks earlier than the current plan.
$60,000
$50,000
$40,000 HOLIDAY.LIGHTS
MAG
$30,000 PRINT
RADIO
$20,000 OOH
$10,000
$-
Week1 Week6 Week11 Week16 Week21 Week26 Week31 Week36 Week41 Week46 Week51
Theme Park: Projecting Results of Optimal Media Spending
To address the stagnant attendance trends for the park, our solution calls for an optimized
spending plan as outlined below. This includes the benefits from optimizing media schedul-
ing. The opportunity points towards a+9.3 percent growth in attendance without the require-
ment of spending incremental marketing funds.
Optimal Spending Strategy estimated
Change in Daily Attendance to result in about a +9.3 percent gain in
daily attendance
400
350 14
300
52
250
200
197
150
100
50 92
0
-50
-40
-100
-26% spend $104K Radio +126% Print adv Media Schedule Marketing
OOH Adv adv spend Optimization Synergy
Conclusions and Recommendations
The results of this modeling exercise has enabled us to crystallize some key insights
regarding key drivers of the Theme Park business.
We learned that the Theme Park’s attendance trends have been flat to stagnant over the past
year. While cooler weather, rising fuel and admissions prices were major negative factors,
total marketing efforts barely covered these short-falls.
We know that temperature and weather has a significant impact on short-term daily
attendance and we succeeded in quantifying those effects.
We uncovered and quantified a new challenge to the park’s efforts to grow attendance, rising
fuel prices. Continued increases in fuel prices are likely to pose a major challenge and risk to
growing the park’s attendance.
We accurately quantified the regular and recurring seasonal and days-of-the-week effects on
park attendance.
We learned that the +5% increase in admissions price at the park cost -3.7 percent in
foregone attendance.
Over the past year, the park’s marketing efforts, while generating greater than break-
even revenues of $1.56 per dollar spent, fell short in terms of driving substantial
growth in park attendance.
With the impact of pricing factored in, the net effect of all media and marketing efforts
generated only a +0.3 percent year-over-year impact.
That there was a -22% decline in the efficiency of media in terms of revenue per expenditure.
A major reason for this decline in efficiency is due to the large reduction in radio advertising.
Going forward, our modeling has uncovered a key opportunity for the Theme Park to
re-ignite attendance growth and improve overall marketing efficiencies.
Our optimized solution calls for a shift in marketing spending by the park
Increase investments in print, radio and Holiday Lights TV advertising
Reduce spending on outdoor and magazine ads
Improve media scheduling by maximizing weeks of activity and minimizing ad hiatus’.
In addition, the park should continue and expand past successful efforts through special
themed events and promotions
By optimizing spending going forward the park can reverse stagnant attendance trends and
grow overall attendance by +9.3 percent, all without the requirement of increasing total
marketing budgets.
Model Validation
The insights presented here are only as good as the models
produced. To assure the highest level of integrity from this
analysis, the following is a summary of our validation efforts.
This validation effort is designed to demonstrate the “predictive
ability” of these models. In order to demonstrate and validate
this validity, we withheld 15% of the data observations
completely from the models. We then compared how well our
models predict actual performance across these unknown data
points. The results of this exercise are shown on the following
chart. In all cases, our holdout forecasts showed high predictive
capabilities and were well within tolerances so we can conclude
that these models are all robust and highly predictive.
Actual v. Model Fit
R2 =96.2%, Holdout R2= 87.9%, MAPE=+/-8.7%
140,000
120,000
100,000
80,000
Actual
60,000 Model
40,000
20,000
0
4 4 4 4 4 4 4 5 5 5 5 5 5 5 5 5 6 6 6 6 6 6 6
.0 9.0 0.0 1.0 2.0 3.0 5.0 6.0 9.0 0.0 1.0 3.0 4.0 5.0 6.0 8.0 8.0 2.0 3.0 4.0 6.0 6.0 8.0
07 1 3 1 2 0 1 2 0 2 0 1 2 0 1 2 0 2 0 1 2 0 1
0 4. 05. 06. 08. 09. 11. 12. 01. 03. 04. 06. 07. 08. 10. 11. 12. 02. 03. 05. 06. 07. 09. 10.
Contact
Michael Wolfe, President
Bottom-Line Analytics LLC
404.841.1620
MJW@bottomlineanalytics.com
www.bottomlineanalytics.com