Traditional market research encompasses various methods and approaches used to gather, analyze, and interpret data about markets, consumers, and competitors before the advent of digital and automated tools. These traditional methods often rely on direct & manual approaches such as surveys, focus group discussions, experiments and test markets and panel studies.
However, above methods comes with limitations for example:
Cost: Some methods, such as in-person surveys and focus groups, can be expensive due to logistical and operational expenses.
Time-Consuming: Collecting and analyzing data through traditional methods can be time-intensive, delaying the availability of insights.
Limited Reach: Methods like mail surveys and telephone interviews may not reach all demographic groups effectively, leading to potential biases.
This brings us to the benefits of utilizing data analysis tools such as Python.
What is Python?
Python is a high-level, versatile programming language known for its readability and simplicity. It is widely used in various fields, including data analysis, web development, automation, and more. Python’s extensive range of libraries and frameworks makes it particularly valuable for complex tasks.
Using python for marketing research
Python provides numerous advantages for market research, enhancing both the efficiency and depth of your analysis. Depending on your proficiency with Python, you can leverage it to automatically collect large datasets, perform data analysis and transformation, and create visualizations that make your insights clearer and more actionable.
Here are some practical examples of how I use Python for market research
You can visit my Github Portfolio section to see some of my Python projects
Utilizing the YouTube Comments Section
Imagine you are a market researcher for a mobile phone distributor, and you need to explore strategies to expand your company’s market share. To achieve this, you aim to understand your competitors’ customers and analyze their sentiments towards both your competitor’s brand and your own.
A valuable, unbiased data source for this analysis is the YouTube comments section. Below, I explain how I utilized Python for market research to gauge iPhone users’ sentiments towards Apple and their opinions on Samsung’s latest flagship product, the Galaxy S23.
TextBlob for NLP of Youtube Comments
Breakdown content into words to identify key terms
Key Takeaways from This Analysis
iPhone users appreciate several features of the Galaxy S23, including the camera, S Pen, screen size, battery life, and gaming capabilities. They particularly value the camera’s advanced zoom capabilities and high-definition shots. For the S Pen, users enjoy the customizable button patterns and its functions for signing documents, doodling, and note-taking. Regarding the screen, they appreciate the responsive touch experience and the immersive, TV-like display.
You can use the above analysis to personalize or tailor your messaging and content when targeting iPhone users to promote Android products.
Comparing Reviews to Products Sold
Imagine you’re responsible for managing your brand’s retail stores on online shopping platforms like Shopee or Lazada. You want to determine if reviews contribute to generating more orders and use this insight to justify a marketing budget for campaigns aimed at increasing product reviews from loyal customers.
Below is how I used Python to quickly analyze the correlation between product reviews and product orders on Shopee.
Load data from Dataframe
Analyse sales per store
Find relationships per variables
Strong correlation between ratings volume vs product sold
Key Takeaways from This Analysis
The graph illustrates a direct linear relationship between the volume of product ratings and the number of items sold on Shopee. The volume of ratings plays a crucial role in driving sales, regardless of the average rating score. Shop owners should consider implementing campaigns or strategies to encourage buyers to leave ratings and reviews, as this can significantly boost sales.
Customer Membership Types
Suppose you are responsible for your company’s customer retention strategy and are seeking methods to enhance engagement, build brand loyalty, and encourage repeat purchases. Using your current customer data, you aim to personalize your loyalty program for future customers based on factors such as age, location, total spend, item purchases, and other relevant data points.
Below, I explain how I utilized Python for market research and developed a model that your marketing team can use to automatically recommend the most suitable loyalty program for newly acquired customers.
Sample data from Kaggle was used, and Python’s Scikit-Learn library employed the K-Nearest Neighbors (KNN) algorithm as the machine learning model.
Data source: https://www.kaggle.com/datasets/uom190346a/e-commerce-customer-behavior-dataset
Load dataset
Perform Exploratory Data Analysis
Start of data modeling using sci-kit learn.
K Nearest Neighbor
Model evaluation with Plot Confusion matrix
Key Takeaways from This Analysis
- Most customers spend less than USD 500. Notably, there are also segments of customers who spend around the median amount, as well as those who exceed USD 1,000. This observation suggests that customers can be categorized into three distinct groups: low spenders, medium spenders, and high spenders.
- There is no clear linear relationship between age and the number of items purchased. This means that older people aren’t necessarily buying more items than younger people, or vice versa.
- By utilizing Age, Total Spend, Item Purchase, and Average Rating as key variables, the KNN model achieves perfect accuracy, with predictions aligning precisely with the original membership types in the test dataset.
In summary, Python is a powerful tool for market research, offering extensive capabilities that streamline and enhance the research process. Its robust libraries enable efficient data collection, sophisticated analysis, and insightful visualization, making it easier to uncover trends, predict future behaviors, and make data-driven decisions.