✅ What is the Primary Purpose of Conjoint Analysis?
Conjoint Analysis is a tool used in marketing to understand what features or aspects of a
product are most important to customers.
🎯 Primary Purpose:
To find out which combination of features (like price, color, size, brand, etc.) customers
like the most — and what they are willing to trade off to get what they want.
📌 Why is it Useful in Marketing?
It helps companies:
● Design better products.
● Set the right price.
● Choose which features to include or remove.
● Understand what matters most to the customer.
🧃 Simple Example:
Imagine you're a company selling fruit juice.
You want to know what your customers prefer. You create a few versions of juice with different
combinations of:
● Flavor: Mango / Orange / Apple
● Sugar Level: No Sugar / Low Sugar / Normal
● Price: ₹20 / ₹30 / ₹40
You show different combinations of these options to people and ask:
"Which one would you buy?"
For example:
● Option 1: Mango, No Sugar, ₹40
● Option 2: Orange, Low Sugar, ₹30
● Option 3: Apple, Normal Sugar, ₹20
Based on their choices, you can figure out:
● Do people care more about price or flavor?
● Are they willing to pay more for no sugar?
● Is Mango more popular than Apple?
✅ Result:
Now, you can create the most liked product — like Mango Juice with Low Sugar at ₹30 —
and market it better.
💡 Summary:
Conjoint Analysis helps businesses understand what customers value most in a product by
testing different combinations of features. It’s like asking people, "If you had to choose, what
would you pick?" — and then using that information to make smarter marketing decisions.
practical examples of how Conjoint Analysis is used in marketing:
✅ 1. Product Design
A mobile phone company wants to know what features customers care about most — like
camera quality, battery life, brand, or screen size — so they can design a phone that people
really want.
✅ 2. Pricing Strategy
A coffee shop uses conjoint analysis to test different combinations of coffee types, sizes, and
prices to find out which price point gives the best balance between customer satisfaction and
profit.
✅ 3. Feature Prioritization
A car manufacturer wants to know whether customers prefer sunroof, automatic gear, or low
fuel consumption — so they include the most desired features in their next model.
✅ 4. New Product Launch
A snack brand wants to launch a new flavor of chips. They use conjoint analysis to test
combinations like spice level, packaging design, and price to choose the best combination for
launch.
✅ 5. Branding Decisions
A toothpaste company wants to know whether customers value brand name, whitening, or
herbal ingredients more. This helps them decide how to position their brand.
✅ 6. Subscription Plans
A streaming service (like Netflix) uses it to understand what customers prefer in their
subscription plans — like HD quality, number of screens, or monthly price — and build better
plans.
✅ 7. Market Segmentation
A travel company analyzes preferences for flight time, luggage allowance, and price. They
group customers based on what matters most to each group and create targeted packages.
✅ 8. Packaging Design
A soft drink brand wants to know whether customers prefer can, bottle, or tetra pack, and
what size they prefer. This helps them design the most appealing packaging.
✅ What is Conjoint Analysis?
Conjoint Analysis is a marketing research method used to understand what features (or
attributes) of a product people care about the most.
🎯 Main Goal:
To figure out how customers make choices between different product options and what
they're willing to trade off — for example, paying more for better quality.
📦 Imagine a Simple Example:
You're selling laptops, and customers can choose based on:
● Brand: Dell, HP
● RAM: 8 GB, 16 GB
● Price: ₹40,000, ₹50,000
You show different combinations (called profiles) and ask people which they prefer. This helps
you understand what matters most: brand, RAM, or price.
📘 Now Let’s Understand the Terms:
I. Part-Worth Functions (Utility Values)
These show how much value or preference customers place on each feature option.
✅ Example:
If customers love 16 GB RAM more than 8 GB, the part-worth (utility) of 16 GB will be higher.
Feature Option Part-Worth
RAM 8 GB +5
RAM 16 GB +15
👉 So, 16 GB adds more "value" in their eyes.
II. Relative Importance Weights
This shows how important each attribute (feature) is compared to others.
✅ Example:
Attribute Relative Importance
RAM 50%
Brand 30%
Price 20%
👉 RAM is the most important factor for customers when choosing a laptop.
III. Attribute Levels
These are the options within each feature.
✅ Example:
Attribute Levels
Brand Dell, HP
RAM 8 GB, 16 GB
Price ₹40,000,
₹50,000
Each level is tested in different combinations.
IV. Pairwise Tables
These are tables where customers are shown two options side-by-side and asked which they
prefer.
✅ Example:
Option A Option B Preferred
Dell, 8 GB, ₹40,000 HP, 16 GB, ₹50,000 Option B
👉 Helps understand trade-offs like “paying more for better RAM.”
V. Cyclical Designs
This is a way to rotate combinations so that each option appears fairly, and you get balanced
comparisons.
✅ Think of it like:
"Mix and match all possible features in a smart way so each one gets tested equally without
repeating the same combination too much."
VI. Fractional Factorial Designs
This is used when testing all combinations is too much.
✅ Example:
If you have 3 features with 3 levels each → 3×3×3 = 27 combinations.
That’s too many.
So, you select a smaller set (like 9) that still gives you meaningful data.
VII. Orthogonal Arrays
These are special designs to ensure combinations are statistically balanced — each level of
each feature appears the same number of times.
✅ This helps make sure the results are fair and unbiased.
VIII. Internal Validity
This checks if the conjoint analysis was done correctly — does the data actually make
sense?
✅ Example:
If a customer always chooses laptops with higher RAM, that makes sense.
If choices are random or contradictory, the study may have low internal validity.
🧠 Final Summary:
Term Meaning (In Simple Words)
Part-Worths How much people like each feature
Importance Weights Which feature matters most overall
Attribute Levels The choices for each feature
Pairwise Tables Show two product options side-by-side
Cyclical Designs Rotate features fairly in tests
Fractional Factorial Use fewer combinations to save time
Designs
Orthogonal Arrays Make sure all features are tested evenly
Internal Validity Check if your results make logical
sense
key steps in conducting Conjoint Analysis, with a clear example for each step.
✅ Overview:
Conjoint analysis is like asking people to choose between different versions of a product,
and then using that data to figure out what features they care about most.
Let’s walk through the process step by step using a simple example:
You're a company launching a new smartphone, and you want to know which features (price,
battery, brand, etc.) are most important to customers.
🔑 Key Steps in Conjoint Analysis
1. Problem Formulation
👉 What do you want to learn?
You define your objective clearly.
✅ Example:
You want to know which smartphone features are most important to customers — brand,
battery life, and price.
2. Stimulus Construction (Creating Product Profiles)
👉 Build sample product combinations with different features.
✅ Example:
Profile Brand Battery Price
A Samsun 4000 mAh ₹15,000
g
B Apple 5000 mAh ₹25,000
C Xiaomi 3000 mAh ₹10,000
These combinations (called stimuli) will be shown to the customer for comparison.
3. Input Data Selection
👉 Collect responses from the target audience.
You ask people to either:
● Rank the profiles
● Rate them (e.g., 1 to 5)
● Choose one out of a pair or a group
✅ Example:
You show customers 2 or 3 profiles and ask:
"Which smartphone would you prefer?"
They choose Profile B → you record that.
4. Conjoint Analysis Model (Data Analysis)
👉 Use software to calculate:
● Part-worths (how much value people give to each feature option)
● Relative importance of each attribute
✅ Example Result:
Feature Level Part-Worth
Brand Apple +20
Samsung +10
Xiaomi 0
Battery 5000 mAh +15
4000 mAh +10
3000 mAh 0
Price ₹10,000 +20
₹15,000 +10
₹25,000 0
👉 Higher part-worth = more preferred.
5. Result Interpretation
👉 Understand what matters most.
✅ Example Interpretation:
● People prefer Apple, but price is even more important.
● Customers prefer lower prices over higher battery or brand name.
This helps you decide:
“Let’s make a low-priced phone with decent battery and maybe a mid-range brand.”
6. Reliability Assessment
👉 Check if the results are trustworthy.
You verify:
● Do the choices make logical sense?
● Do people consistently choose better options?
● Can the results be repeated?
✅ Example:
If someone chooses an expensive phone with poor battery over a cheaper, better one, it might
be a red flag.
You check if overall patterns are consistent and reliable.
✅ Summary Table:
Step What It Means Simple Example
1. Problem Define the goal "What features do people want in a
Formulation phone?"
2. Stimulus Create feature combinations Different smartphone profiles
Construction
3. Input Data Collect choices from people Customers pick favorite phone
Selection
4. Conjoint Model Analyze the choices Get part-worth values
5. Interpretation Understand what's most Price matters more than brand
important
6. Reliability Check if results make sense Are answers consistent?
✅ What is the Conceptual Basis of Factor Analysis?
Factor Analysis is a statistical method used to find hidden patterns or groups in a large set
of variables.
🎯 Concept in Simple Words:
It helps us reduce many variables (questions, features, etc.) into a few underlying factors
that explain most of the data.
🧠 Simple Example:
Imagine you did a survey asking 10 questions about a person's shopping habits, like:
● I compare prices before buying.
● I wait for discounts.
● I shop when there's a sale.
● I read online reviews.
● I care about brand image.
● I only buy trusted brands.
● I avoid new or unknown brands.
● I read product descriptions carefully.
● I always look at product ratings.
● I shop online more than offline.
Now, instead of analyzing 10 separate things, Factor Analysis may find:
● Factor 1: Price Consciousness (questions 1–3)
● Factor 2: Brand Preference (questions 5–7)
● Factor 3: Information-Seeking Behavior (questions 4, 8–10)
So we simplified 10 questions into 3 main ideas.
📊 What Type of Data Does It Require?
● It needs quantitative data — typically collected from surveys or questionnaires with
responses on a Likert scale (e.g., 1 to 5: Strongly Disagree to Strongly Agree).
● The data must have some correlation between variables. If all variables are totally
unrelated, factor analysis won’t work.
🔍 What Does PCA Stand For?
PCA = Principal Component Analysis
It is a technique similar to factor analysis, often used to reduce the number of variables while
keeping the maximum possible information.
✅ In simple terms:
PCA transforms many variables into a smaller set of “principal components” that still capture
most of the variation in the data.
🔁 PCA is often the first step before doing factor analysis.
🏷️ What is the Purpose of Naming Factors?
Once factors are created (like Factor 1, Factor 2…), we need to give them meaningful names
based on the variables they include.
🎯 Purpose:
● Makes interpretation easier.
● Helps businesses or researchers understand what each factor represents.
● Allows us to take action based on grouped insights.
✅ Example:
If Factor 1 includes:
● I wait for discounts.
● I compare prices.
● I shop during sales.
You might name this factor: “Price Sensitivity”
🧠 Quick Summary Table:
Concept Simple Explanation Example
Factor A method to group related variables into hidden 10 shopping questions → 3
Analysis patterns (factors) factors
Data Numeric data from surveys with Likert scales 1–5 scale questions
Needed
PCA Principal Component Analysis — reduces data Turns 10 variables into 3
while keeping key information components
Naming Giving a clear label to each group of related Group about sales →
Factors questions "Price Sensitivity"
