toy horse conjoint analysis
TRANSCRIPT
Toy Horse Conjoint Analysis Leonardo Zhang
Executive Summary Key Points and Recommendations: u Through bene8it segmentation we found 3 consumer segments and the product No. 15, No. 13 and No. 2 are best target to each segment.
u Through priori segment level conjoint analysis we con8irmed and quanti8ied age/
gender differences in styling preferences. No. 12 meets the overall group’s needs, gender differences are more signi8icant than age.
u Through simulate market shares and pro8its for different product-‐line scenarios we suggest EarlyRiders modify existing No. 13 product to No. 15, drop No. 5 product and add No. 2 product.
Taken competitor’s reaction into consideration. A new product with No. 15 and No. 2 products will generate the highest pro8it ($158,333, 82.58% increase compared with statue quo $86,657). By then the market share of EarlyRiders will increase from 36.4% to 65%. u Combined two segmentation analyses together we found No. 15 is very suitable for targeting girls and 1~2 year-‐old group, and product No. 2 is more preferred by boys and 3~4 year-‐old group. We recommend EarlyRiders tailor packing images, locations for ads and promotion events accordingly.
1: BeneJit Segmentation via Cluster Analysis of Conjoint Part-‐Utilities Test cluster approaches and select the best – 3 clusters u Approach 1: “toclust” function (result: 3 clusters) toclust = partworths; pm1 = pamk(toclust,scaling=TRUE) pm1$nc u Approach 2: use sum of squares plot (result: 3 clusters) Test 2, 3 and 4 clusters 3 clusters 8its best
Segmentation output
u Segment size
Constant Price Height Motion Style
u Segment center data
u Segment description and targeted product segment 1: Prefer: Price=119.99, Height=26’, Motion=rocking, Style=glamorous. (“height” is the most important attribute) Product that best targets this segment: Product ID 15 $139.99+26’+rocking+glamorous (Meet all their preferences except price, not very price sensitive) segment 2: Prefer: P=119.99, H=18’, M=rocking, S=glamorous. (“price” and “motion” are the most important attributes, style is the least in8luential attribute) Product that best targets this segment: Product ID 13 $139.99+18’+rocking+glamorous (Meet all their preferences except price, not very price sensitive) segment 3: Prefer: P=119.99, H=26’, M=bouncing, S=racing. (“price” is the most important attribute, “height” is the least in8luential attribute) Product that best targets this segment: Product ID 2 $119.99+18’+bouncing+racing (Meet all their preferences except height, since they are very price sensitive we set a lower price. Consider our cost and the less in8luential of height, we will give them 18’)
segment segment segment
2: Priori segment level conjoint analysis Overall regression result without segment
"price" is the most in8luential attribute "style" attribute is the least important In average, moving from 0 to 1 on "price", "height” and "style" attributes increase utility, on "motion" decrease utility overall preferred feature: “price”=119.99, “height”=26', “motion”=bouncing, “style”=glamorous (Product ID 12)
Price Height Motion Style
Segment by age
Only "height" and "motion” have signi8icant difference. Means 1~2 year-‐old age group has different part-‐worth in H and M attributes compared to overall group. 1~2 year-‐old group is more prefer “height”=26 and “motion”=bouncing. 3~4 year-‐old group is less sensitive with “height”.
Overall 1~2 year-‐old
3~4 year-‐old
Segment by gender
All attributes are signi8icant. Means female group has different part-‐worth in all attributes compared to overall group. Female group more prefer "height"=26, "motion"=rocking and "style"=glamorous, and less sensitive in “price”. Male group is more price sensitive, and prefer "motion"=bouncing and "style"=racing.
Overall Female
Male
EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
13&5 7 0.364 0.636 86,657 180,950
Status Quo
Scenario 1: Add No. 15 & No. 2 Because No. 13, No. 15 & No. 2 best target 3 segments we analyzed before. EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
13&5&15&2 7 0.979 0.021 223,148 13,364
Scenario 2: Add No. 15 & No. 2, Drop No. 5 Because No. 5 is less appealing, when provided with No. 13 & No. 7 at the same time, all 3 segments will not select No. 5. EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
13&15&2 7 0.977 0.023 240,708 -‐12,732
Scenario 2’: Consider competitor lower price However, in Scenario 2, the competitor will try to decrease price and win back market share and pro8it, so our actual pro8it will be $142,733. EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
13&15&2 8 0.662 0.338 142,733 54,346
3: Simulate market shares & proJits for different product-‐line scenarios
Scenario 3 and Scenario 4: Choose between 15 and 13
15&2 7 0.973 0.027 256,725 -‐11,469 15&2 8 0.65 0.35 158,222 56,986 13&2 7 0.758 0.242 193,609 56,462 13&2 8 0.462 0.538 99,637 98,338
EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
After take a deeper look at the products combination in Scenario 2, we 8ind that 15 and 13 are similar and may cause cannibalization. Thus, we test the scenarios of only producing 1 product of those two. Still in both scenarios, competitor will lower price for it’s own interest.
As a result, When producing 15 & 2 and competitor products 8, EarlyRider will yield a pro8it of $158,222, which is higher than that when producing 13 & 2.
Scenario 5: Treat the market as a whole? If we treat the market as a whole, the most favorable for all people is product 12 ($119.99, 26”, Rocking, Glamour), so we test the scenario of only produce product 5. Also competitor in this scenario will lower his price.
EarlyRider's product ID
Competitor's product ID
EarlyRider's market share
Competitor's market share
EarlyRider's pro8it ($)
Competitor's pro8it ($)
12 7 0.928 0.072 228,666 2,749 12 8 0.596 0.404 139,704 68,863
To conclude, EarlyRide should develop a product line with produce No. 15 & No. 2
No.15: $139.99, 26”, Rocking, glamorous No. 2: $119.99, 18”, Bouncing, Racing
According priori segment level conjoint analysis by gender in Part 2, No. 15 is very suitable for targeting girls. And product No. 2 is more preferred by boys. Also since 1~2 year-‐old group is more height sensitive, No. 15 product is more suitable for it and No. 2 product is more suitable for 3~4 year-‐old group.
Implications for the company's marketing plans
Price Height Motion Style
Female Less price sensitive 26’ Rocking Glamorous
No. 15 $139.99 26’ Rocking Glamorous
Price Height Motion Style
Male More price sensitive Less height sensitive Bouncing Racing
No. 2 $119.99 18’ Bouncing Racing
4: Concrete Recommendations to EarlyRiders
Use different marketing approaches for No. 15 and No. 2 to better targeting different groups. Avoid cannibalization and gain each consumer segments’ market shares as more as possible. For No. 15 product: u Targeting to girls and 1~2 year-‐old group u Design feminine packing images and product appearance u Tailor product design better meet 1~2 year-‐old group’s 8igure u Choose toddler zone in toy stores/ TV shows and cartoons targeting girls for ads u Promotion event idea: different customs for toy house as gift u Emphasize product functions in marketing For No. 2 product: u Targeting to boys and 3~4 year-‐old group u Design masculine packing images and product appearance u Tailor product design better meet 3~4 year-‐old group’s 8igure u Choose 3~4 year-‐old kids zone in toy stores/ TV shows and cartoons targeting boys for ads
u Promotion event idea: toy horse racing competition, collaborate with kindergartens u Emphasize low price and boy preferred features in marketing
Appendix
Plot (2 clusters) Plot (4 clusters)
Not chosen segmentation output
GenderD AgeD
1 female 1~2 year-‐old
0 male 3~4 year-‐old
Values of dummy variables used in regressions
Product ID Price Height Motion Style 1 0 0 0 0 2 1 0 0 0 3 0 1 0 0 4 1 1 0 0 5 0 0 1 0 6 1 0 1 0 7 0 1 1 0 8 1 1 1 0 9 0 0 0 1 10 1 0 0 1 11 0 1 0 1 12 1 1 0 1 13 0 0 1 1 14 1 0 1 1 15 0 1 1 1 16 1 1 1 1
Price 1 = 119.99 0=139.99 Height 1=26" 0=18" Motion 1=Rocking 0=Bouncing
Styles 1=Glamorous 0=Racing
Scenario 2 utility simulation
No. 13 Product
No. 5 Product
No. 7 Product
Segment 1 61 50 66 Segment 2 47 46 39 Segment 3 13 19 24
Simulation Code