<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Muwahiddin Massaid on Medium]]></title>
        <description><![CDATA[Stories by Muwahiddin Massaid on Medium]]></description>
        <link>https://medium.com/@mmassaid?source=rss-19ddcd31e563------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*t4iVa0_6O2Q0tnSLRGTAKQ.jpeg</url>
            <title>Stories by Muwahiddin Massaid on Medium</title>
            <link>https://medium.com/@mmassaid?source=rss-19ddcd31e563------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 05:01:26 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mmassaid/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[CROSS-SELLING IN INSURANCE COMPANY: WHO IS INTERESTED IN VEHICLE INSURANCE?]]></title>
            <link>https://medium.com/@mmassaid/cross-selling-in-insurance-company-who-is-interested-in-vehicle-insurance-562aba225bf8?source=rss-19ddcd31e563------2</link>
            <guid isPermaLink="false">https://medium.com/p/562aba225bf8</guid>
            <category><![CDATA[classification-models]]></category>
            <category><![CDATA[classification]]></category>
            <category><![CDATA[predictive-analytics]]></category>
            <category><![CDATA[predictions]]></category>
            <category><![CDATA[classification-algorithms]]></category>
            <dc:creator><![CDATA[Muwahiddin Massaid]]></dc:creator>
            <pubDate>Sat, 09 Jul 2022 05:20:04 GMT</pubDate>
            <atom:updated>2022-07-16T11:38:47.685Z</atom:updated>
            <content:encoded><![CDATA[<p>Predicting which customers who paid for health insurance are also interested in vehicle insurance.</p><p><strong>SUMMARY</strong></p><p>This study aims to predict customers’ interest using data from <a href="https://www.kaggle.com/datasets/anmolkumar/health-insurance-cross-sell-prediction?datasetId=869050&amp;sortBy=voteCount&amp;select=train.csv"><strong>Kaggle.</strong></a> Exploration data found that the age range of 36 to 55 years old is the most enthusiast age in vehicle insurance, even though most of the data are under 30 years old. Those who had vehicle damage experienced more interest than no. Also, having paid annual premiums in the range of 20.000–40.000 shows more interest.</p><p><strong>INTRODUCTION</strong></p><p><a href="https://www.simon-kucher.com/en-us/about/media-center/simon-kucher-partners-report-explores-untapped-potential-cross-selling-insurance-us">A study by consulting firm Simon-Kucher &amp; Partners</a> shows that only 4–9% of U.S customers insurance (survey respondents) confirmed that they were contacted about additional coverage even though more than 84% of them said they have been in contact with their insurer within the last year.</p><p>Why are so few customers offered additional insurance products?</p><p>It is said that <a href="https://insuranceagentapp.com/cross-selling-insurance-current-clients/#:~:text=Cross%2Dselling%20insurance%20offers%20several,overall%20revenues%20in%20the%20process.">cross-selling makes customers can think that the company only wants more money</a> so, some agencies respond by lower their cross-selling activities.</p><p>In fact, cross-selling is an effective technique to boost company revenue. Selling to existing customers is more cost-effective than acquiring new customers. It is known that <a href="https://customerthink.com/5-ways-marketers-should-be-measuring-success/">80% of future profit comes from 20% of existing customers.</a></p><p>One right step to cross-selling is to predict whether customers are interested or not in an additional product. It can optimize the company’s chance to offer the right product to the right customers so, they will think the company can understand their needs.</p><p>This article shows predicting customers’ interest in vehicle insurance from the Kaggle dataset <a href="https://www.kaggle.com/datasets/anmolkumar/health-insurance-cross-sell-prediction?datasetId=869050&amp;sortBy=voteCount&amp;select=train.csv"><strong>here</strong></a>. Analysis performed using Python <a href="https://github.com/massaid01/VehicleInsurance/blob/main/Prediction_of_Interest_in_Vehicle_Insurance.ipynb"><strong>here</strong></a> and model machine learning deployed via Streamlit web app <a href="https://massaid01-vehicleinsurance-streamlitapp-2-cyiwm0.streamlitapp.com/"><strong>here</strong></a>.</p><p>This article provides the following questions</p><ul><li>Which features show customers’ interest in vehicle insurance?</li><li>How can models of machine learning help this cross-selling process?</li><li>Are there any features more influence than others on customers’ interest?</li></ul><p><strong>DATA FEATURES</strong></p><p>The following are the features in the data source to be researched:</p><ul><li><strong>Response (Target Variable)</strong>: Interested or Not</li><li><strong>Gender</strong>: Customer gender</li><li><strong>Age</strong>: Customer age</li><li><strong>Driving_License</strong>: Customer already has DL or not</li><li><strong>Region_Code</strong>: Unique code for the customer region</li><li><strong>Previously_Insured</strong>: Customer already has vehicle insurance or not</li><li><strong>Vehicle_Age</strong>: Vehicle age</li><li><strong>Vehicle_Damage</strong>: Customer got his/her vehicle damaged in the past or not</li><li><strong>Annual_Premium</strong>: The amount customer needs to pay as a premium in the year</li><li><strong>Policy_Sales_Channel</strong>: Anonymized code for the channel of outreaching to the customer ex: Different Agents, Over Mail, Over Phone, In Person, etc.</li><li><strong>Vintage</strong>: Number of days customer has been associated with the company</li></ul><p><strong>FINDINGS</strong></p><blockquote><strong>Which features show customers’ interest in vehicle insurance?</strong></blockquote><p>To answer this question, graphs are used to see the correlation between each feature and target variable (Response Customers).</p><p><strong>Age feature</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*V3kqTxFuJUYFfzxxtE_9Tg.png" /><figcaption>Histogram for age features</figcaption></figure><p>Based on the graph, most of the data age is under 30 years old. This may be a trend in the area where the young are more aware of the importance of financial products than the older. To get more insight, we grouped them into ranges age.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zhjwsJQXlTUxOY39IkOVOw.png" /><figcaption>(Left) Age features, (Right) Age features with conversion rate</figcaption></figure><p>Based on the graph, the age group of 36 to 55 years is the most favored vehicle insurance, although most of the data are under 30 years old. This may happen because of better financial conditions to buy more comfortable living facilities.</p><p><strong>Vehicle Damage Feature</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*amyq5mvZRIzL0ueDkb2A1w.png" /><figcaption>(Left) Vehicle damage features, (Right) Vehicle damage features with conversion rate</figcaption></figure><p>Customers who had vehicle damage experienced are more interested than those who have never yet. It is a good insight for the company to offer them vehicle insurance.</p><p><strong>Annual Premium</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Fp0SN6aCkd_06QDOOMgpBg.png" /><figcaption>(Left) Range of annual premium features, (Right) Range of annual premium features with conversion rate</figcaption></figure><p>Customers paying annual premiums in the range of 20.000–40.000 are the most interested group in vehicle insurance. It is a sign to company to dig deeper about offer type they give in 20.000–40.000 annual premium in order to more understanding customers character.</p><blockquote><strong><em>How can machine learning models help this cross-selling process?</em></strong></blockquote><p>Machine learning can used to predict customer interest in vehicle insurance so that companies have a higher chance of cross-selling to existing customers appropriately.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ZJZGPcPOFfArwwRMzZO5MA.png" /><figcaption>Recreating a table of some machine learning models. You can see the details process <a href="https://github.com/massaid01/VehicleInsurance/blob/main/Prediction_of_Interest_in_Vehicle_Insurance.ipynb">here</a></figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pIFNq01KSVqUZUcagqH02w.png" /><figcaption>Recreating a table of LGBM model scores. You can see the details process <a href="https://github.com/massaid01/VehicleInsurance/blob/main/Prediction_of_Interest_in_Vehicle_Insurance.ipynb">here</a></figcaption></figure><p>Based on the graph, <em>Light gradient boosting machine (LGBM)</em> is the most suitable models for predict this dataset. Within AUC score 0,86 on test data and 0,85 on test data. This means that the model is accepted.</p><p>What does it mean with an 0,85 (85%) AUC score?</p><p>Here is the interpretation</p><ul><li>This model can predict 85% of interested customers. This means that if there are 1,000 data and 120 data are predicted to be interested in vehicle insurance, then 102 (85% of 120) may really be interested in your offer.</li><li>Assuming the annual premium for vehicle insurance is 500, then you have the potential additional revenue from this cross-selling is 51,000 from existing customers without incurring additional marketing costs. It is effective &amp; efficient too because you only approach those who are interested.</li></ul><blockquote><strong>Are there any features more influence than others on customers’ interest?</strong></blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*S4OG8OAQY-WECYTG1F0GXQ.png" /><figcaption>Featured importance</figcaption></figure><p>Based on the chosen machine learning model, there are some features more affect the target variable than others. <em>Vintage, Region_Code, Annual_Premium </em>and<em> Age</em> are the most influence features than others.</p><p><strong>Conclusion &amp; Recommendation</strong></p><ul><li>Since most customers are under 30 years old, the company needs some strategy to nurture customers as long as the company can. In fact, the range age of 35 to 55 is the most enthusiast in vehicle insurance.</li><li>Promoting vehicle insurance to customers who had past vehicle damage experienced</li><li>Research more about annual premiums in the 20,000 to 40,000 range to gain more insight into why customers with these annual premiums tend to prefer vehicle insurance offers.</li><li>If possible, the company can provide information about the type of vehicle and financial measures such as job customers. This can improve machine learning prediction accuracy.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=562aba225bf8" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>