![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2022\/05\/pexels-monicore-134065-1024x538.jpg\")
<\/p>\n<\/p>\n
Churn prediction – tandem with engagement – is probably the most wanted use case we get from Marketing departments across industries. For those of you that do not know what churn is, basically it\u2019s associated with customers that will leave your company\/services. So, it shouldn\u2019t be a surprise that companies put a lot of effort into making sure once a customer is acquired, you don\u2019t lose it.<\/span><\/p>\n We can think of churn from multiple angles, but all of them are solved by – more or less – the same analytical models. For instance:<\/span><\/p>\n Master the art of transforming companies with AI.<\/p>\n\t\t\tSaber mais<\/a>\n\t\t<\/div>\n\t<\/div><\/span><\/p>\n In this section, we will cover churn from multiple perspectives, including some insights from our previous experience on how to boost these models.<\/span><\/p>\n To be realistic, all customers are churners. Eventually, they will stop buying your services (voluntary churn) or you will shut down your company\/product line (involuntary churn). So, when we think about predicting churn as a binary signal, it is restricted to a certain observance window. For instance, who is churning in the next week, in the next month, in the next quarter, etc? The predictive function, in this case, tends to assume this form:<\/span><\/p>\n <\/span> <\/span> Models that predict churn at a more immediate future tend to be more accurate. Namely, it\u2019s easier to identify unsatisfied customers that will initiate a contract termination tomorrow than a year from now. However, short-term churn prediction is less actionable since, for most of those customers, the damage may be irreparable. So, we should aim for a right trade-off between prediction accuracy and recoverability\/actionability.<\/span><\/p>\n In general, this category falls into binary classification, which allows us to use any classification model we want as long as it\u2019s capable of dealing with extremely unbalanced data in the order of 1-99 or even less than that. Yes, it is bad news for Data Scientists, but good news for the Business side – otherwise your business would bankrupt soon. Even if we consider the target as a binary label, consider the model outcome as a probability and treat those with a probability high enough to compensate for the treatment costs. So, favor false positives since a lost client is hard – or even impossible – to recover. False negatives aren\u2019t that bad unless you\u2019re offering major discounts\/benefits for re-engaging.<\/span><\/p>\n The main disadvantage of this approach is that it doesn\u2019t give you an urgency indicator of when the customer will leave, nor they tell you what actions would be required to heal him.<\/span><\/p>\n I\u2019ve seen people using SHAP values<\/a> to \u201cdecide\u201d how to save them. Please, don\u2019t go this way. It is simply wrong. If your model thinks a customer will churn because he hasn\u2019t used your services for a month, the right treatment isn\u2019t inviting him to get into your website. Keep in mind your models are learning correlations and not causality.<\/span><\/p>\n Time-to-churn models aim to predict when the customer is leaving the company. These models can be based on standard regression, ordinal classification with time segmentation into classes, or using survival analysis prediction. The predictive function, in this case, tends to assume this form:<\/span><\/p>\n <\/span> <\/span> The main ambiguity of time-to-event modes is how to handle customers that do not churn. Namely, how to include active customers during training? You can use semi-supervised learning (assuming you don\u2019t know their label) or weakly supervised learning (regularizing the prediction to be higher than your observance window without explicitly saying the actual value).<\/span><\/p>\n <\/p>\n Once you know the time to churn, you can schedule clients to handle first the customers that will leave the company sooner.<\/span><\/p>\n While these models are more informative regarding urgency than binary classification ones, they still don\u2019t tell you how to save the customer.<\/span><\/p>\n From our experience, the right way to model churn is using uplift modeling. At uplift modeling, you aim to discover the effect of a marketing action on a customer. For instance, how the probability of churn decreases given a discount, marketing campaign, etc. While learning in a counterfactual manner is difficult – i.e., you don\u2019t know what would have happened if you didn\u2019t do the action, you can estimate the effect by applying different actions to different customers and learning the probability of churn conditioned on the action. The predictive function, in this case, tends to assume this form:<\/span><\/p>\n <\/span> <\/span> where F is any of the two methods previously referred to (churn as binary classification or time to churn) but conditioned on the action. For the sake of simplicity, let\u2019s assume you\u2019re modeling F as a binary classification task. In this case, the model is telling you the expected benefit from applying an action. Therefore, you are no longer targeting your more unsatisfied customers but you\u2019re targeting overall loyalty. What\u2019s the action that I should apply to each and every customer to engage him?<\/span><\/p>\n As for modeling actions into your model, don\u2019t go with abstract actions. Try to be as discriminative as possible in the features. Include features such as: what\u2019s the contact channel? what\u2019s the day of month\/week\/time of day? what’s the discount percentage I\u2019m offering? What are the additional perks I\u2019m including? In this way, your model will be able to generalize for new actions you do in the future.<\/span><\/p>\n Of course, actions have cost. So, you probably don\u2019t want just the action that minimizes churn at any expense (which would likely be paying your customer for using your service). You want the action that has the best trade-off between retaining the customer and making some profit out of it. Let\u2019s say the predicted revenue for an alive customer is R(customer) (refer to our previous blog post<\/a> to learn how to estimate this<\/span>).<\/span><\/p>\n Also, we have a function that tells us the cost of an action C(action). Assuming a user that leaves the service generates $0 (it may not be the case since interrupting contracts tend to incur on expenses on both sides), the final function that tells you the best action per client is:<\/span><\/p>\n <\/span> <\/span> In case the best action has a negative expected gain, the best thing is to let the customer go. Otherwise, go for that action and keep both your customer and CFO happy.<\/span><\/p>\n We have built dozens of churn models for multiple industries<\/a>. Therefore, if reducing churn* is one of your goals as a company and you\u2019re looking at predictive analytics, let\u2019s have a call and discuss how we can collaborate.<\/span><\/p>\n * increasing engagement, upselling, cross-selling, etc.<\/span><\/i><\/p>\n \n\n \n
\n\t\t![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2022\/08\/4-194x301@3x.png\")
<\/div>\n\t\tCourse, Templates<\/h6>
Data Ignite<\/h3>\n\t\t\t
How to model churn?<\/span><\/h2>\n
Churn as a binary classification<\/span><\/h3>\n
![\"](\"https:\/\/nilg.ai\/wp-content\/ql-cache\/quicklatex.com-e4a3a94c291fc394be60e4d32d245b95_l3.png\" "\\"Rendered")
<\/p>\n![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2021\/01\/Time-vs.-Accuracy-vs.-Actionability-1.svg\")
<\/a><\/p>\n![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2021\/01\/Confusion-Matrix.svg\")
<\/a><\/p>\nTime to churn<\/span><\/h3>\n
![\"](\"https:\/\/nilg.ai\/wp-content\/ql-cache\/quicklatex.com-271d564b024e8963061e0bd672db0bec_l3.png\" "\\"Rendered")
<\/p>\n![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2021\/01\/Survival-Analysis.svg\")
<\/a><\/p>\nUplift modeling<\/span><\/h3>\n
![\"](\"https:\/\/nilg.ai\/wp-content\/ql-cache\/quicklatex.com-85893945188551184386b4319511185c_l3.png\" "\\"Rendered")
<\/p>\n![\"](\"https:\/\/nilg.ai\/wp-content\/ql-cache\/quicklatex.com-b4816776ecddb684be08208d56505b82_l3.png\" "\\"Rendered")
<\/p>\n![\"\"](\"https:\/\/nilg.ai\/wp-content\/uploads\/2021\/01\/Uplift.svg\")
<\/a><\/p>\n\n\t\t\n\t\t \n\t\t ![\"Kelwin](\"https:\/\/nilg.ai\/wp-content\/uploads\/2022\/08\/Web-Kelwin.png\")
\t\t <\/div>\n\n