extraalearn-customer-conversion-classification

Context

Over the last decade, the EdTech sector has experienced tremendous growth. It's projected that by 2023, the online education market could reach a valuation of $286.62 billion, expanding at a compound annual growth rate of 10.26% from 2018 to 2023. The shift towards online education has been significantly driven by its numerous advantages, such as easy access to information, personalized learning paths, and transparent assessments, making it a preferred choice over traditional educational methods.

The COVID-19 pandemic has further accelerated the expansion of the online education industry, attracting many new entrants and customers. The widespread availability and ease of digital marketing tools allow companies to effectively reach a broader audience. Potential customers who express interest in these educational offerings are referred to as leads. EdTech companies typically generate leads through various channels including:

• Interaction with marketing content on social media or other online platforms.
• Website or app visits where the customer might download informational brochures.
• Email inquiries for more detailed information.

Companies nurture these leads with the goal of converting them into paying customers through direct communications via phone calls or emails.

Objective

ExtraaLearn, an emerging startup, offers cutting-edge technology courses to both students and professionals aiming to enhance their skills. The startup faces the challenge of determining which leads are most likely to convert to paying customers amidst the high volume of leads it generates. As a data scientist at ExtraaLearn, you are tasked with:

• Developing a machine learning model to predict lead conversion likelihood.
• Identifying key factors that influence the conversion of leads.
• Creating a profile for leads that are most likely to convert.

Data Description

The dataset includes various attributes of leads and their interactions with ExtraaLearn, detailed as follows:

Data Dictionary

• ID: Unique identifier for each lead.

• Age: Age of the lead.

• Current Occupation: Current job status of the lead, with possible values including 'Professional', 'Unemployed', and 'Student'.

• First Interaction: Initial contact method with ExtraaLearn, options being 'Website' or 'Mobile App'.

• Profile Completed: The extent of profile completion by the lead on the platform, categorized as Low (0-50%), Medium (50-75%), or High (75-100%).

• Website Visits: Count of the lead's visits to the website.

• Time Spent on Website: Total duration of all visits to the website by the lead.

• Page Views per Visit: Average number of pages viewed during a visit.

• Last Activity: Most recent interaction of the lead with ExtraaLearn.

• Email Activity: Includes activities like requesting program details or receiving informational brochures via email.

• Phone Activity: Includes phone conversations or SMS interactions with a representative.

• Website Activity: Includes actions like live chatting with a representative or updating the profile on the website.

• Print Media Type1: Indicates whether the lead saw an ExtraaLearn ad in the newspaper.

• Print Media Type2: Indicates whether the lead saw an ExtraaLearn ad in a magazine.

• Digital Media: Indicates exposure to ExtraaLearn ads on digital platforms.

• Educational Channels: Indicates whether the lead discovered ExtraaLearn through educational channels like forums, educational websites, etc.

• Referral: Indicates if the lead came through a referral.

• Status: Indicates if the lead converted to a paid customer.

Actionable Insights and Recommendations

Insights from Logistic Regression

• Analysis using logistic regression reveals that referrals, the website being the first point of contact, and recent activities are key predictors of customer conversion.
• ExtraaLearn should particularly focus on professionals as they are more likely to become paying customers.
• Data indicates that customers who have not completed their profiles extensively are less likely to convert. Thus, prioritizing those with complete profiles could be beneficial.
• Interestingly, we found that engagements over the phone tend to have a lower conversion rate. It might be worth reevaluating phone-based outreach strategies as potential customers could perceive them as intrusive.
• Our findings suggest that unemployed individuals and students have lower conversion rates. Adjusting pricing strategies or offering tailored promotions might make courses more accessible to these groups.
• Enhancing the referral program could be advantageous, as referred leads show a higher likelihood of converting.

Insights from Decision Trees

• The decision tree model identifies website interactions, the duration of website visits, and profile completion rates as crucial factors in customer conversion.
• Effective lead qualification should focus on three main criteria:
  • The customer initiated contact through the website rather than being approached by cold calls.
  • The customer completed a significant portion of their profile.
  • The customer spent substantial time on the website, ideally more than three hours.

These tailored strategies based on logistic regression and decision tree analyses can help ExtraaLearn more effectively convert leads into loyal customers.

Recommendations for Business

Enhance Digital Touchpoints and User Experience

1. Optimize the Website Experience: Ensure that the website is user-friendly, with clear calls to action and easy navigation. This will encourage visitors to spend more time and engage more deeply, increasing the likelihood of conversion.
2. Improve Profile Completion: Develop strategies to encourage complete profile completion. This could include prompts during the user journey, offering incentives for completing profiles, or highlighting the benefits of a full profile.

Tailor Marketing and Sales Strategies

3. Refine Targeting of Professionals: Since professionals have a higher conversion rate, tailor marketing efforts to attract this demographic. Utilize professional networking sites and platforms for targeted advertisements.
4. Adjust Communication Tactics: Given the negative reception to phone calls, consider shifting focus to less intrusive communication methods like email or messaging through the platform. This approach can be personalized and less disruptive.
5. Develop Special Offers for Students and Unemployed: Introduce discounts, flexible payment plans, or scholarships specifically designed for students and unemployed individuals to make learning more accessible and boost conversion rates in these segments.

Leverage Data and Analytics

6. Enhance the Referral Program: Strengthen the referral program by offering current customers incentives for referring new students. Clearly communicate the benefits of the referral program through all channels.
7. Utilize Analytics for Lead Scoring: Implement a lead scoring system based on the insights from the decision tree analysis. Prioritize leads who meet the identified criteria (website initiation, high profile completion, and extensive site engagement).

Foster Engagement and Build Relationships

8. Interactive Content and Webinars: Host webinars and interactive content sessions that require registration. This can help in collecting data and also keeps the user engaged with the platform, thereby increasing the chances of conversion.
9. Community Building: Foster a sense of community among users by encouraging interaction through forums, user groups, or social media groups. This can help maintain engagement and potentially lead to higher conversion rates.

Monitor and Adapt Strategies

10. Continuous Monitoring and A/B Testing: Regularly track the performance of these strategies and conduct A/B testing to understand what works best. Use data-driven insights to continuously refine approaches to lead conversion.