Product Sales
Background
Pens and Printers was founded in 1984 and provides high-quality office products to large organizations. The company is a trusted provider of everything from pens and notebooks to desk chairs and monitors. Although it does not produce its own products, it sells items made by other companies.
Over the years, Pens and Printers has built long-lasting relationships with its customers, who rely on the company to provide them with the best products to meet their needs. As consumer purchasing behavior evolves, the company’s sales strategies must adapt as well. Launching a new product line is a significant investment, and it is essential to use the most effective techniques to ensure its success. Since the most effective approach may differ for each new product, the company must quickly identify what works and what does not.
Six weeks ago, Pens and Printers launched a new line of office stationery. Despite the world becoming increasingly digital, there remains a strong demand for notebooks, pens, and sticky notes.
The company’s focus has been on offering products that help customers be more creative, with an emphasis on tools that support brainstorming. To promote this new product line, three different sales strategies were tested: targeted emails, phone calls, and a combination of both.
-
Email: Customers in this group received an email at the time of the product launch and a follow-up email three weeks later. This approach required minimal effort from the sales team.
-
Call: Customers in this group were contacted directly by a member of the sales team. On average, each call lasted approximately thirty minutes per customer.
-
Email and Call: Customers in this group were first sent an email with information about the product line. One week later, they received a follow-up phone call from the sales team to discuss their needs and how the new products could support their work. While the email required little effort, the phone calls averaged around ten minutes per customer.
Dataset Overview
The dataset consists of 15,000 records, capturing key details of customer transactions related to product sales. It is structured to provide insights into customer behavior, sales performance, and the effectiveness of different sales methods.
Dataset Features:
- Customer ID: A unique identifier for each customer, allowing for tracking of repeat purchases and customer behavior over time.
- Week: The specific week in which the transaction occurred, useful for analyzing sales trends over time.
- Number of Products Sold (nb_sold): Represents the quantity of items purchased in a single transaction.
- Revenue: The total dollar amount generated from the sale, providing insight into average spending per transaction.
- Sales Method: The approach used to engage with the customer before purchase. The dataset includes three methods:
- Email: Customers were contacted via email.
- Call: Customers were approached through phone calls.
- Email + Call: A combination of both methods was used.
- Number of Site Visits (nb_site_visits): The number of times a customer visited the website before making a purchase, serving as an indicator of engagement and purchase intent.
- State: The geographical location of the customer, useful for regional sales analysis and understanding location-based sales trends.
- Years as a Customer: The length of time a customer has been associated with the company, helping to analyze the relationship between customer tenure and purchase behavior.
Purpose of the Dataset Analysis
This dataset was analyzed to:
- Evaluate customer purchasing patterns and sales performance.
- Assess the impact of different sales methods on revenue.
- Understand customer engagement levels through site visits and transaction history.
- Identify trends based on customer location and years of association with the company.
- Provide data-driven recommendations to optimize sales strategies and revenue growth.
This structured dataset serves as a foundation for uncovering actionable insights, helping businesses refine their sales approaches and improve overall customer engagement.
Data Validation
The dataset contains 15000 rows and 8 columns before cleaning and validation. I have validated all the columns against the criteria in the dataset description.
- week: Numeric values without missing values, same as the description. No cleaning is needed.
- sales_method: 5 models before cleaning, 2 of which were 'em + call' and 'email' that were replaced with 'Email + Call' and 'Email' respectively to match the description.
- customer_id: 15000 unique values, same as the description. No cleaning is needed.
- nb_sold: Numeric values without missing values, same as the description. No cleaning is needed.
- revenue: 1074 missing values were detected. They were replaced with the overall median revenue, 89.50.
- years_as_customer: No missing values but some abnormal values were detected. These abnormal values were greater than the number of years the company has been in existence for (39), indicating that these customers had been buying from the company before it was founded, which is not possible. I assumed this was a random error indicating customers that had been buying from the company since inception. I replaced them with the number of years the company has been in existence (39).
- nb_site_visits: Numeric values, same as the description. No cleaning is needed.
- state: 50 unique values, indicating the number of states in the United States without missing values. No cleaning is needed.
After the data validation, the dataset contains 15000 rows and 8 columns without missing values.
How many customers were there for each approach?
From the last 6 weeks' record, majority of the customers (approximately half of the 15000) were approached by Email. The second most used approach was the Call method, of which approximately 5000 out of 15000 customers were approached by. The Email and Call method was the least employed method out of the three with about 2600 customers reached through this approach.
What does the spread of the revenue look like overall?
The revenue from sales made per customer, regardless of the approach used, ranges from around 30 to 240 USD. The overall median revenue per customer is around 90 USD. The average sale made to a customer, regardless of approach used, is just about 94 USD. 75 percent of unit sales (sales made per customer) was above 53 USD.
What does the spread of the revenue look like for each method?
Email Method: The revenue from sales made per customer approached by this method ranges from around 80 to 150 USD. Half of the customers approached by this method purchased just about 94 USD worth of products. The average sale made to each customer approached by this method is approximately 97 USD.
Email and Call Method: The Email and Call Method has the largest range of revenue per customer with majority of customers approached this way spending above 150 USD, which is the highest spent by any customer approached by either the Email or Call method. The customer with the highest purchase was also contacted using the Email and Call Method. The median revenue made per customer was 182 USD. The average sale made per customer approached this way is just around 170 USD.
Call Method: The revenue from sales made via this approach ranges around 30 to 90 USD. The majority of sales made through this approach is within 42 and 53 USD. 50 percent of sales were below 50 USD. The maximum sale made through this approach was around 90 USD, which is the minimum sale made via the Email and Call Method and just above the minimum through the Email Method.
The Email and Call Method is the least used approach by a sizeable margin. This needs to change. The Call Method should be discontinued. It takes the most time and energy from the sales team as each call takes thirty minutes on average. I highly recommend that the Email and Call Method be used much more because it makes the company the most revenue on average and does not require as much effort as the Call Method.
Was there any difference in revenue over time for each of the methods?
For the customers approached by the Email Method, the company gathers most of its revenue via this approach within the first two weeks of product launch, after which the sales begin to dwindle.
The sales for the customers approached by the Email and Call Method is slow at the beginning but eventually picks up from weeks four to six.
Regarding the customers approached by the Call Method, the sales experienced a steady increase for the first five weeks after the product launch.
We need to dig deeper into why the customers approached through the Email and Call Method are late bloomers, meaning they do not start purchasing in bulk until the third week after product launch.