{"id":1039,"date":"2023-05-03T09:57:57","date_gmt":"2023-05-03T09:57:57","guid":{"rendered":"https:\/\/artelnics.com\/?post_type=case_studies&#038;p=1039"},"modified":"2026-11-09T11:38:09","modified_gmt":"2026-11-09T11:38:09","slug":"telecommunications-sector-case","status":"publish","type":"case_studies","link":"https:\/\/artelnics.com\/case_studies\/telecommunications-sector-case\/","title":{"rendered":"Preventing churn in the telecommunications industry"},"content":{"rendered":"<div class=\"aui aui-content-page artelnics-shared-page\" data-artelnics-shared-content>\n<article class=\"artelnics-shared-case\">\n<p class=\"artelnics-shared-case-lead\">A predictive model identifies telecommunications customers at risk of leaving so retention teams can prioritize action.<\/p>\n<div class=\"artelnics-shared-case-meta\"><span class=\"artelnics-shared-case-tag\">Telecommunications<\/span><span class=\"artelnics-shared-case-tag\">Churn prediction<\/span><span class=\"artelnics-shared-case-tag\">Customer retention<\/span><\/div>\n<h2>Challenge<\/h2>\n<p>A telecommunications company needed to identify customers likely to leave for a competitor. Churn reduces future revenue, removes opportunities to offer additional services and can affect brand reputation.<\/p>\n<h2>Data<\/h2>\n<p>The dataset contained fewer customers who left than customers who stayed, creating a class imbalance. This can cause conventional models to favor the majority class and overlook customers at risk of leaving.<\/p>\n<p>Individual variables showed low logistic correlations with churn, below 10%, so predicting churn required a model that considered multiple factors together.<\/p>\n<h2>Solution<\/h2>\n<p>Artelnics developed a predictive model using an algorithm that minimizes weighted quadratic error to address the class imbalance.<\/p>\n<p>The evaluation considered sensitivity and specificity alongside the other classification measures. For an imbalanced dataset, apparently strong overall measures can conceal poor detection of customers who actually leave.<\/p>\n<div class=\"aui-card-grid aui-grid-two\">\n<div class=\"aui-card\"><strong>Method<\/strong><\/p>\n<p>Predictive modeling with weighted quadratic error.<\/p>\n<\/div>\n<div class=\"aui-card\"><strong>Output<\/strong><\/p>\n<p>Identification of customers at high risk of churn.<\/p>\n<\/div>\n<\/div>\n<h2>Results<\/h2>\n<p>The reported classification test gave 67.80% precision, 32.20% error, 74.47% sensitivity and 67.59% specificity. The area under the ROC curve was 0.78.<\/p>\n<table>\n<thead>\n<tr>\n<th colspan=\"2\">Classification test<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Precision<\/td>\n<td>67.80%<\/td>\n<\/tr>\n<tr>\n<td>Error<\/td>\n<td>32.20%<\/td>\n<\/tr>\n<tr>\n<td>Sensitivity<\/td>\n<td>74.47%<\/td>\n<\/tr>\n<tr>\n<td>Specificity<\/td>\n<td>67.59%<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The model helps retention teams identify high-risk customers and prioritize opportunities to prevent churn. Some customers flagged as high risk may remain, so predictions should support retention decisions rather than be treated as certain outcomes.<\/p>\n<h2>Illustrations<\/h2>\n<figure><img decoding=\"async\" src=\"https:\/\/artelnics.com\/wp-content\/uploads\/2023\/08\/logistic-correlation-telecom.png\" alt=\"Logistic correlation between the variables\" \/><figcaption>Logistic correlation between the variables<\/figcaption><\/figure>\n<figure><img decoding=\"async\" src=\"https:\/\/artelnics.com\/wp-content\/uploads\/2023\/08\/ROC_curve_Telecom.png\" alt=\"ROC curve churn prevention\" \/><figcaption>ROC curve churn prevention<\/figcaption><\/figure>\n<\/article>\n<nav class=\"artelnics-shared-related\" aria-label=\"Case study actions\">\n<div><strong>Have a similar challenge?<\/strong><span>Artelnics can help turn operational data into a practical AI system.<\/span><\/div>\n<div class=\"aui-actions artelnics-shared-related-actions\"><a class=\"aui-button aui-button--secondary\" href=\"\/case_studies\/\">View all projects<\/a><a class=\"aui-button\" href=\"\/contact\/\">Talk to us<\/a><\/div>\n<\/nav>\n\n\n<\/div>","protected":false},"excerpt":{"rendered":"<p>A predictive model identifies telecommunications customers at risk of leaving so retention teams can prioritize action.<\/p>\n","protected":false},"featured_media":622,"parent":0,"menu_order":0,"template":"","categories":[],"class_list":["post-1039","case_studies","type-case_studies","status-publish","has-post-thumbnail","hentry"],"acf":[],"_links":{"self":[{"href":"https:\/\/artelnics.com\/api\/wp\/v2\/case_studies\/1039","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/artelnics.com\/api\/wp\/v2\/case_studies"}],"about":[{"href":"https:\/\/artelnics.com\/api\/wp\/v2\/types\/case_studies"}],"version-history":[{"count":14,"href":"https:\/\/artelnics.com\/api\/wp\/v2\/case_studies\/1039\/revisions"}],"predecessor-version":[{"id":3372,"href":"https:\/\/artelnics.com\/api\/wp\/v2\/case_studies\/1039\/revisions\/3372"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/artelnics.com\/api\/wp\/v2\/media\/622"}],"wp:attachment":[{"href":"https:\/\/artelnics.com\/api\/wp\/v2\/media?parent=1039"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/artelnics.com\/api\/wp\/v2\/categories?post=1039"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}