Business Forecasting - 4th Sem MBA Notes

Prof. Chitranka K, ch.E, chart PR, ChartMktr

Module 3 - Market Research

Marketing Research Techniques

1. Qualitative vs. Quantitative: Choose between qualitative methods (interviews,

focus groups) for in-depth insights and quantitative methods (surveys, analytics)
for statistical data.
2. Desk Research: Utilize existing data from reports, studies, and databases to
inform your understanding without primary data collection.
3. Field Research: Collect data through observations, interviews, and experiments in
real-world settings to gain fresh insights.
4. Observational Studies: Analyze consumer behavior in natural settings to
understand actions and preferences without direct interaction.
5. Experimental Research: Test hypotheses by manipulating variables in controlled
settings to establish cause-and-effect relationships.
6. Social Media Listening: Monitor online conversations and consumer sentiments
about brands and products to gain real-time insights.
7. Competitive Analysis: Study competitors' strategies, strengths, and weaknesses
to identify opportunities and threats in the market.
8. Ethnographic Research: Immerse in the consumer's environment to observe
behaviors and contexts influencing buying decisions.
9. Longitudinal Studies: Track changes over time in consumer attitudes and
behaviors to identify trends and shifts in preferences.
10. Content Analysis: Systematically analyze textual or visual communication (like
social media posts) to understand key themes and messages.

Consumer Surveys

1. Clear Objectives: Define specific goals to determine what information the survey
needs to capture.
2. Target Audience: Identify and select the right demographic to ensure relevant
data collection.
 3. Question Design: Craft clear, unbiased questions with appropriate response
options (e.g., multiple choice, Likert scale) to minimize confusion.
4. Mix of Question Types: Use a combination of closed (quantitative) and
open-ended (qualitative) questions for comprehensive insights.
5. Pretesting: Conduct pilot surveys to identify and rectify potential issues in
question clarity or survey flow.
6. Anonymity Considerations: Assure participants of their anonymity to encourage
honest and accurate responses.
7. Length and Timing: Keep surveys reasonably short to maintain engagement and
avoid participant fatigue.
8. Distribution Channels: Utilize various platforms (online, email, in-person) based
on where the target audience is most accessible.
9. Incentives: Offer rewards or incentives to boost participation rates and gather a
larger sample size.
10. Data Analysis: Use statistical tools to analyze and interpret survey data,
identifying significant trends and correlations.

Consumer Clinics

1. Product Testing: Allow consumers to interact with a product prototype or concept

and provide feedback on their experience.
2. Interactive Sessions: Facilitate direct engagement between consumers and
product developers to encourage dialogue and detailed feedback.
3. Observation: Observe how participants use the product to uncover usability
issues and pain points.
4. Diverse Participants: Recruit a diverse group representative of the target market
to obtain a wide range of perspectives.
5. Facilitated Discussions: Guide conversations that explore consumer perceptions
and experiences in depth.
6. Feedback Mechanisms: Utilize structured forms or discussions to gather specific
feedback on features, design, and overall satisfaction.
7. Use of Prototyping: Present prototypes or mockups during the clinic to gauge
initial reactions and suggestions for improvement.
8. Iterative Testing: Implement feedback into product redesigns, allowing for
multiple rounds of testing to refine the product further.
9. Follow-Up: Seek follow-up interviews or surveys to see if suggestions made
during the clinic were effective in the final product.
10. Documentation: Record sessions (with consent) to capture nuanced feedback
and ideas for improvement.
Focus Groups

1. Structured Discussions: Facilitate group discussions around specific topics,

products, or brand messages to gather collective insights.
2. Moderator Role: Employ a skilled moderator to guide discussion, encourage
participation, and keep the conversation on track.
3. Participant Diversity: Assemble groups that reflect different demographics,
attitudes, and experiences to ensure varied input.
4. Open-Ended Questions: Use open-ended questions to stimulate conversation and
gather in-depth responses from participants.
5. Group Dynamics: Leverage the interaction among participants to uncover deeper
insights and varying perspectives.
6. Recording and Transcribing: Capture discussions through audio or video
recordings for accurate analysis and insights extraction.
7. Comfortable Environment: Create a relaxed atmosphere to encourage honest
sharing and active participation from all members.
8. Debriefing Sessions: Conduct post-focus group discussions with the team to
analyze findings and brainstorm implications.
9. Incorporating Nonverbal Cues: Pay attention to body language and interaction
styles for additional context beyond verbal responses.
10. Synthesis of Feedback: Summarize key themes and insights from the
discussions to inform marketing strategy and product development.

Market Experiments in Test Stores

Objective:
Market experiments (also known as controlled experiments) are conducted to evaluate
the effects of marketing interventions or operational changes in real-time. Test stores
serve as laboratories for these experiments.

Design Steps:

● Selection of Test vs. Control Stores:

Randomly choose stores to implement the intervention (test stores) and others
to serve as a benchmark (control stores). This selection should account for
similar historical sales patterns and characteristics to ensure comparability.
 ● Implementation of Interventions:
It can include new promotions, pricing strategies, product placements, or
changes in customer service protocols.
● Data Collection Pre- and Post-Experiment:
Gather sales and customer interaction data before and after the intervention to
assess its impact.
● Evaluation:
Analyze differences in performance metrics between test and control stores,
using statistical methods to infer causality from observed differences.

Statistical Estimations

Definition:
Statistical estimation involves making inferences about a population based on sample
data. The goal is to accurately estimate key parameters.

Key Concepts:

● Point Estimation:
This provides a single value estimate of a parameter (e.g., average sales) using
statistics such as the sample mean.
● Interval Estimation:
Offers a range within which the parameter is expected to lie, typically expressed
as a confidence interval (e.g., a 95% confidence interval for the mean).
● Methods Used:
Depending on the nature of the data and distribution, various methods (like
Maximum Likelihood Estimation) can be applied.

Variable Identification

Process:
 ● Independent Variables (Predictors):
These are the variables that you manipulate or observe to see their effect on the
dependent variable (e.g., advertising spend, price discounts).
● Dependent Variable (Outcome):
This is the primary variable of interest that you are trying to predict or explain
(e.g., total sales).
● Control Variables:
These are additional variables that, while not the primary focus, can influence the
relationship and need to be controlled for to avoid bias (e.g., seasonality,
competitor pricing).

Time Series and Cross-Sectional Data Collection

Time Series Data:

● Definition: Data points collected or recorded at specific time intervals (e.g., daily,
weekly, monthly).
● Characteristics: Commonly shows trends (long-term movements), seasonality
(repeating patterns), and cyclic behavior.
● Applications: Analyzing sales trends to forecast future performance or to assess
impacts of marketing campaigns over time.

Cross-Sectional Data:

● Definition: Data collected at a single point in time across multiple subjects or

entities (e.g., survey results from various stores on the same date).
● Characteristics: Useful for comparing different units at a given time.
● Applications: Assessing variations in customer satisfaction across multiple
locations.
Specification of the Model

Model Formation:

● Selecting Relationships: Determine how different variables are expected to

interact (e.g., a linear relationship between sales and advertising expense).
● Model Types: Choose from various modeling approaches such as:
○ Linear Regression: Models a straight-line relationship.
○ Logistic Regression: For binary outcomes (e.g., purchase: yes/no).
○ Polynomial Regression: For relationships that may follow a curved line.

Functional Form:
Denotes how variables relate mathematically. The choice can impact model accuracy
significantly. For non-linear relationships, transformations such as logarithmic or
polynomial forms might be more appropriate.

Estimation of the Parameters

Parameter Estimation Techniques:

● Ordinary Least Squares (OLS):

The most common method for estimating the parameters of a linear regression.
It minimizes the sum of squared differences between observed and predicted
values.
● Assumptions:
OLS relies on several assumptions:
○ Linearity: The relationship between independent and dependent variables
is linear.
○ Independence: Observations are independent of each other.
○ Homoscedasticity: Constant variance of the error terms.
Interpretation of Regression Statistics

Once parameters are estimated, interpreting those results is crucial for drawing
conclusions:

● Coefficients:
Each coefficient measures the change in the dependent variable for a one-unit
change in the independent variable, holding all other variables constant.
● R-squared Value:
Indicates how well the model explains the variability of the dependent variable. A
higher R-squared suggests a better fit.
● P-values:
Used to determine the statistical significance of each coefficient. A small p-value
(typically < 0.05) indicates that the null hypothesis can be rejected, suggesting a
significant relationship between the predictor and the outcome.

Time Series Regression

Purpose:
Used when the dependent variable is a time series, allowing for more targeted
forecasting and analysis of temporal patterns.

● Components of Time Series Regression:

○ Trend Analysis: Captures long-term movement in the data.
○ Seasonality: Identifies repeating patterns based on time frames (e.g.,
holidays).
○ Cyclic Behavior: Accounts for fluctuations affecting the data over longer
periods.
● Stationarity Testing:
Before performing a time series analysis, it's essential to check if the series is
 stationary (mean, variance do not change over time). Techniques like the
Augmented Dickey-Fuller test can be used. If not stationary, differencing or
transformation methods may be needed.

Forecasting with Regression Models

Forecasting is a critical application of regression analysis. Different methods can be

employed based on the data properties and specific needs:

● Unconditional Forecasting:
Uses the fitted regression model to predict future values without adjusting for
previous errors or values. This is straightforward and best for stable
environments.
● Forecasting with Serially Correlated Errors:
In situations where residuals are correlated, adjusting the model (e.g., using
autoregressive integrated moving average (ARIMA) models) or including lagged
variables can improve accuracy.
● Conditional Forecasting:
Generates predictions based on certain assumptions or conditions (like
macroeconomic indicators or expected changes in market conditions). This
approach allows decision-makers to assess various scenarios.

Possible Questions
1. Explain the difference between qualitative and quantitative research methods.
Provide an example of each.
2. What are the key steps involved in conducting consumer clinics?
3. Describe the importance of pretesting surveys before full deployment.
4. What are the assumptions of Ordinary Least Squares (OLS) regression?
5. Define point estimation and interval estimation in the context of statistical
estimation.