Section 1: Understanding the Need for MDA
Introduction
In today’s competitive smartphone market, customer preferences are shaped by a variety
of factors such as price, service, features, and brand value. For DigiTech, which has been
facing declining sales despite offering consistent products, it became important to truly
understand what drives customer satisfaction. Traditionally, companies rely on univariate
and bivariate analysis to make decisions. While these methods are simple, they are not
always sufficient in markets where customer behavior is influenced by multiple
interrelated variables. This is exactly the situation DigiTech faced — the company
needed to move beyond simple averages and two-variable comparisons to a more
advanced method, namely Multivariate Data Analysis (MDA).
1. Why were DigiTech's univariate and bivariate analyses insufficient in identifying the
true drivers of customer satisfaction?
Univariate analysis studies one variable at a time. For example, DigiTech looked at the
average satisfaction score of 7.2 or the average battery rating of 8.0. Although these
numbers look positive, they do not tell us why some customers are more satisfied than
others. The limitation here is that it provides only surface-level information. In a highly
competitive market, making strategic decisions solely on averages can mislead managers.
For instance, just knowing that the battery rating is good may tempt the company to keep
improving batteries, but that does not guarantee higher satisfaction.
Bivariate analysis, on the other hand, looks at the relationship between two variables at a
time. For DigiTech, it showed that customers with higher satisfaction also gave slightly
higher battery scores, while dissatisfied customers complained about after-sales service.
This approach does give more insight compared to univariate methods, but it still fails to
capture the combined influence of multiple variables. For example, a customer may be
satisfied not just because of battery but because of battery plus service quality plus brand
reputation. Since bivariate analysis studies only two variables together, it ignores this
combined effect.
In DigiTech’s case, customer satisfaction is not shaped by just one or two features but by
a mix of many factors. Univariate and bivariate methods were too simplistic to handle
this complexity. This is why they were insufficient in identifying the true drivers of
customer satisfaction.
2. How does multivariate data analysis provide a more realistic understanding of
customer behaviour in a competitive market?
Multivariate Data Analysis (MDA) is designed for situations where multiple factors
interact at the same time. Unlike univariate and bivariate methods, MDA can consider
customer demographics, purchase behavior, and perceptions together. This provides a
holistic picture of customer decision-making.
For example, suppose DigiTech wants to know why one group of customers is satisfied
and another is not. MDA can analyze multiple features like after-sales service, camera,
price sensitivity, and loyalty scores simultaneously. The regression model they used
showed that after-sales service (0.4) had a stronger impact on satisfaction than battery
(0.3) or camera (0.2). This means customers care more about being supported after
purchase than about just technical features. Such insights are only possible through
MDA.
In real-world markets, decisions are rarely made in isolation. Customers consider a
bundle of features before making a purchase. MDA reflects this reality by allowing
businesses to understand how different factors interact and which ones matter the most.
This makes it far more reliable for decision-making than traditional methods.
3. In what ways can univariate or bivariate analysis lead to misleading marketing
decisions?
One major risk of relying on univariate analysis is that it can lead to overemphasis on a
single variable. For DigiTech, if they had focused only on improving battery because it
scored 8.0, they might have ignored service quality — which turned out to be the most
important driver of satisfaction. Such a decision would mean high investment in
specifications but no improvement in customer loyalty or repeat purchases.
Similarly, bivariate analysis can also mislead by suggesting false relationships. For
example, it might show that satisfaction is higher among men compared to women. But
what if the difference is actually due to income level, or the type of phone used, or the
frequency of purchase? By ignoring other variables, bivariate analysis can misattribute
the reason behind satisfaction.
In short, univariate and bivariate analyses give a fragmented view. They can lead
managers to invest in areas that look important on the surface but do not truly drive
customer behavior. DigiTech’s challenge illustrates this perfectly — they risked focusing
on battery improvements when the real issue was after-sales service.
Conclusion
Section 1 highlights why DigiTech’s traditional methods failed and why MDA
became essential. In the smartphone market, where choices are shaped by multiple
interacting factors, simple averages or two-variable correlations cannot capture
reality. Univariate and bivariate analyses are useful for descriptive purposes but
dangerous when used as the sole basis for strategy. Multivariate analysis, on the other
hand, provides deeper insights into what really drives satisfaction. For DigiTech, this
shift from basic analysis to MDA was not just a technical choice but a strategic
necessity for survival in a highly competitive market.
Section 2: Application of MDA Techniques
Introduction
After realizing that traditional methods were insufficient, DigiTech turned to Multivariate
Data Analysis (MDA). This approach allowed them to not only understand the key
drivers of satisfaction but also to segment customers, simplify complex data, and predict
loyalty patterns. In this section, we explore the application of different MDA techniques
regression, cluster analysis, factor analysis, and discriminant analysis and how they can
guide DigiTech’s marketing and operational decisions.
4. Interpret the Regression equation:
Satisfaction = 0.3 (Battery) + 0.4 (After-sales) + 0.2 (Camera) + Error.
What actionable insights can be drawn from this result?
This means that after-sales service contributes the most (0.4), followed by battery
(0.3), and then camera (0.2). The actionable insight here is that customers value
service reliability more than just specifications. For DigiTech, this is a crucial finding
because most smartphone brands compete by advertising technical features like
megapixels or battery hours, while service often remains under-communicated.
DigiTech should therefore strengthen service centers, offer extended warranties, and
highlight customer support in its campaigns. Maintaining good battery and camera
quality is still necessary, but service should be the central focus.
5. How would Cluster Analysis help DigiTech design better marketing campaigns?
Cluster analysis helps group customers into segments with similar characteristics. For
DigiTech, such segmentation can reveal distinct groups like:
Price-sensitive users: Focused mainly on affordability and battery life.
Tech-savvy youth: Interested in high-performance and advanced camera quality.
Loyal, long-term customers: Value service support and reliability more than flashy
specifications.
Once these clusters are identified, DigiTech can create targeted campaigns. For
example, student clusters may be offered discounts and camera-centric ads, while
older professionals may be attracted with reliability and service guarantees. This
makes marketing more efficient by ensuring that each segment gets a message
tailored to its priorities.
6. What is the role of Factor Analysis in simplifying customer perception variables like
camera, battery, display, and performance?
Customers often evaluate multiple features such as battery, display, camera, speed,
and service. Looking at each separately makes analysis complicated and confusing.
Factor analysis helps by combining correlated variables into fewer underlying
dimensions. For instance, camera, display, and performance might form a single
factor called “Technical Excellence,” while service and battery together might form
“User Reliability.”
For DigiTech, this simplification means managers can focus on two or three broad
factors rather than being lost in ten different metrics. It also helps in designing
strategies that improve overall customer perception instead of only tweaking
individual features.
7. How can Discriminant Analysis be used to predict customer loyalty?
Discriminant analysis allows DigiTech to predict which customers are likely to
remain loyal. By analyzing attributes such as satisfaction, purchase frequency, and
after-sales ratings, DigiTech can classify customers into “loyal” or “at-risk”
categories. This predictive ability is extremely valuable in today’s competitive market
because retaining existing customers is often cheaper than acquiring new ones.
For example, if a customer has average satisfaction but low service ratings, the model
may predict they are at risk of switching. DigiTech can then intervene with loyalty
rewards, special service offers, or personal follow-ups to prevent churn.
Conclusion
The application of MDA techniques provides DigiTech with actionable insights at
multiple levels. Regression shows which features matter most, cluster analysis guides
targeted marketing, factor analysis simplifies perception into meaningful groups, and
discriminant analysis predicts loyalty behavior. Collectively, these techniques allow
DigiTech to go beyond descriptive statistics and make data-driven, strategic choices.
In a market where customers are spoilt for choice, MDA ensures that the company
understands its customers better and invests resources in the areas that truly matter.