Predict Customer Lifetime Value Presentation

Oct 7, 20192 likes1,608 views

This document discusses customer lifetime value (CLV), which is the net present value of future cash flows from a customer. It notes that not all customers are equally profitable and that customer retention can be cheaper than acquisition. Common CLV modeling approaches discussed include RFM analysis, the Pareto/BG-NBD model, and random forest models using features like customer engagement levels. Case studies on Groupon and ASOS are provided that used random forests and neural networks to predict churn and CLV. While neural networks can improve predictions, random forests often provide good results at a lower cost.

A BIRD’S EYE VIEW OF CUSTOMER
LIFETIME VALUE

Contents
1. Why?
2. RFM & BG/NBD (2005)
3. Groupon: Churn + Random Forest (2016)
4. Asos: Random Forest + Embeddings (2017)

Customer Lifetime Value
Customer Lifetime Value (CLTV/CLV/LTV) is the net
present value of the future (net) cash flows associated
with a particular customer.
Peter Fader, Customer Centricity, p 72, Wharton Digital Press, 2012

Why Use Customer LTV?
● It’s not (always) an aggregate metric
● Not all customers are equally
profitable
● Cost of acquisition can be 5x cost of
retention
● What about the time horizon?
○ 1 month future LTV?
○ 3 month future LTV?
○ 1 year future LTV?
○ 1 decade future LTV?
○ 5 decades future LTV?

Business Context
Contractual Non-Contractual
● Customer “death” can be observed
● Often modeled using survival-based
approaches
● Customer “death” is unobserved
● Customer lifetime distribution often
modeled via exponential models
● Amazon Prime
● Spotify Premium
● Netflix
● Gym membership
● Health Insurance
● Playstation+
● Buying stuff from Amazon
● Buying stuff on Ebay
● Buying stuff from Tesco
● Buying video games on Steam
● Buying top ups on Ding

R F M LTV
5 5 4
2 2 3
1 2 5
● Recency
○ How long ago did the customer last
transact?
● Frequency
○ How many times have they transacted?
● Monetary Value
○ How much revenue did they generate?
Pareto (BG)/NBD
Fader et al, 2005

Pareto BG/NBD
Assumptions
● P(Alive | RFM, Timeframe)
● E[Transactions | RFM, Timeframe]

Pareto BG/NBD
- Decent benchmark
- Not appropriate for contractual
settings
- Not appropriate when
transactions can only incur at
fixed intervals (eg festivals)

Pareto Distribution
https://www.kaggle.com/c/ga-customer-revenue-prediction/

Groupon
- 2 stage random forest: Churn + CLV
- 40 features including
where ɑ and β are historical conversion rates for opened and clicked emails
- User segments
(never purchased | new | one-time | sporadic users | power users)
Vanderveld et al, 2016

ASOS
Chamberlain et al, Customer Lifetime Prediction Using Embeddings, KDD, 2017
AUC Churn: 0.79
Spearman CLV: 0.56

ASOS -- Cost
Maximum Area Under the receiver operating characteristics
Curve (AUC) achieved on a test set of 50,000 customers in
hybrid models against the number of neurons in the hidden
layers (in log scale).
Mean monetary cost to train hybrid models on a training set
of 100,000 customers against the number of neurons in the
hidden layers (both in log scale).

Conclusions
● Revenue distribution is not particularly friendly for making
predictions
● RFM & BG/NBD is a good start & benchmark
● Random Forest models probably deliver the most bang for the
buck
● Neural Nets can improve predictions, but at a significant cost

References
● Peter Fader talk on Pareto BG/NBD
● Counting Your Customers the Easy Way [pdf]
● Python Lifetimes Library
● Etsy CLV [pdf]
● 2006 review by Sunil Gupta et al
● An Engagement-Based Customer Lifetime Value System for E-commerce [pdf]
● Customer Lifetime Value Prediction Using Embeddings [pdf]
● Implementing and Training Predictive Customer Lifetime Value Models in Python
● Write-up of the link above

Thank you
Eric Mehes
@meHessian
https://www.linkedin.com/in/eric-mehes/

This document discusses predicting customer lifetime value (CLV) for an auto insurance company using various machine learning models. It outlines preprocessing the data, feature engineering, building models like linear regression, ridge regression, decision trees, random forests, and evaluating the models using R2. Random forest was found to have the best performance with an R2 of 0.913. The conclusions recommend focusing on male customers, specific coverage offers, and tailored sales strategies based on customer demographics to improve CLV predictions.

Customer lifetime value (1)KRISTINA FISCHEROVA

This document discusses customer lifetime value (CLV) and how to optimize marketing strategies around it. It covers: 1. Choosing the right key performance indicators (KPIs) to focus on like CPA, ROAS, profit, or CLV. 2. An overview of customer centricity and how CLV is the "Holy Grail" metric for understanding customer strategies and analytics. 3. How to take action and put the chosen KPI into practice through campaign reporting, optimization, and ensuring the right data integration and technology platforms. 4. A short introduction to marketing experiments for proving the effectiveness and value of marketing efforts through controlled experimentation.

Customer Life Time ValueYunusTalipEROL

Customer Lifetime ValuePaul Woods

This presentation provides insight into how to forecast and calculate customer lifetime value (CLV). Here a startup applied a scientific approach to maximise customer retention and minimise churn. The outputs of the analytics were built into the system and business processes driving the success of the company and helping it to win the customer service of the year award, and to achieve a successful exit through acquisition.

Customer Lifetime ValueEd Kless

This document discusses calculating and analyzing customer lifetime value (CLV). It provides formulas for calculating CLV based on yearly gross contribution per customer, retention rate, discount rate, and retention costs. An example calculation is shown. The document also discusses how retention rate affects CLV and introduces the concepts of net promoter score (NPS) and customer likeability score to categorize customers. It proposes mapping customers on a "grid" based on these metrics to identify the most and least valuable relationships.

customer lifetime valueR S Raghav

Customer lifetime value (CLV) is an important marketing metric that represents the net profit attributed to the entire future relationship with a customer. CLV is calculated as the present value of future cash flows from a customer over their lifetime, taking into account factors like future purchases, retention rates, and costs. Companies are increasingly segmenting customers based on CLV to focus marketing efforts on the most profitable segments. Strategies to increase CLV include upselling, cross-selling additional products to existing customers, and improving customer satisfaction to gain referrals which lowers acquisition costs over the long run.

4.6 customer lifetime valueRick Rasmussen

This presentation discusses customer lifetime value (CLV) modeling. CLV is defined as the net profit attributed to the entire future relationship with a customer. It is typically used in business-to-consumer contexts but can also apply to business-to-business. CLV must be greater than customer acquisition costs. The presentation outlines different levels of sophistication in CLV prediction models and lists key inputs like contribution margin, churn rate, retention cost, and period. It provides a simple e-commerce example and discusses how churn rate strongly impacts CLV for subscription businesses. Standard and more accurate CLV calculation formulas are also shown.

Everything You Need to Know About Customer Lifetime Value (CLV)Demac Media

Customer Lifetime Value (CLV) has become a must know term for eCommerce merchants of any size. It allows you to discover the true value each customer has to your store. But what most people don't know are the different ways to calculate it and the true impact it has on your business. In this discussion we cover everything you need to know about customer lifetime value. We present a few ways to calculate it as well as show some cases where CLV can impact a business. These cases will show what happens when you ignore lifetime value, and the benefits of making decisions with CLV in mind.

Customer Lifetime ValueJY Chun

Presentation on Customer Lifetime ValueKisorkumar Varadarajan

This document summarizes the work of a team that conducted customer segmentation, lifetime value modeling, and recommendations based on the analysis. The team segmented customers using RFM analysis and descriptor variables, identified 7 segments with different characteristics and churn rates. They then modeled customer future value and found segments with high, medium, and low future values. The recommendations target migrating customers between segments to increase revenues.

Customer lifetime valueyalisassoon

The document discusses customer lifetime value (CLV), including: 1. CLV is the predicted net profit attributed to the entire future relationship with a customer. 2. CLV is an important metric because the best customers account for the majority of sales, and retaining existing customers is often cheaper than acquiring new ones. 3. Companies can use CLV to inform customer acquisition and relationship management strategies, such as spending more to acquire or retain the most valuable customers.

Churn Prediction in PracticeBigData Republic

BigData Republic teamed up with VodafoneZiggo and hosted an meetup on churn prediction. Telecom companies like VodafoneZiggo have long benefited from the fine art/science of predicting churn. Currently, in the booming age of subscription based business models (e.g. Netflix, Spotify, HelloFresh), the importance of predicting churn has become widespread. During this event, VodafoneZiggo shared some of its wisdom with the public, after which BDR Data Scientist Tom de Ruijter presented an overview of the modeling tools at hand, both classical, as well as novel approaches. Finally, the participants engaged in a hands-on session showcasing the implementation of different approaches. PART 1 — Churn Prediction in Practice by Florian Maas At VodafoneZiggo we are incredibly excited about Advanced Analytics and the enormous potential for progress and innovation. In our state of the art open source platform we store the tremendous amount of data that is generated every single second in our mobile and fixed networks. This means that we have a vast body of rich information, which if unlocked, can lead to something very special. As a company with a primarily subscription-based service model, churn plays a vital role in the daily business. Not only is the churn rate a good indicator of customer (dis)satisfaction, it is also one out of two factors that determines the steady-state level of active customers. During this talk, we will show how data science provides added value in the process of churn prevention at VodafoneZiggo. We will talk about the data and the modeling approach we use, and the pitfalls and shortcomings that we have encountered while building the model. We will also briefly discuss potential improvements to the current approach, which brings us to talk #2. PART 2 — The Churn Prediction Toolbox by Tom de Ruijter The second talk will show you the fine intricacies of predicting churn through different approaches. We’ll start off with an overview of different modeling strategies for describing the problem of churn, both in terms of a classification problem as well as a regression problem. Secondly, Tom will give you insights in how you evaluate a churn model in a way such that business stakeholders know how to act upon the model results. Finally, we’ll work towards the hands-on session demonstrating different model approaches for churn prediction, ranging from classical time series prediction to recurrent neural networks.

Customer Churn Prevention Powerpoint Presentation SlidesSlideTeam

Introducing Customer Churn Prevention Powerpoint Presentation Slides. Discuss various ways through which a company can manage customer churn with this PPT slide deck. Showcase methods and ways by which a company can prevent the customer from reducing their purchase of products and services. Our readily available PPT slide deck helps to present the types of customer churn, methods to handle customer attrition, the impact of successful implementation of churn management, dashboard, churn propensity model, etc. Take the assistance of customer churn management PPT slideshow to depict several ways by which a firm can experience customer churn such as when customers stop spending, churn due to product quality, etc. Showcase four stages of customer churn management which allow the company to handle customer attrition. Present how the firm can prevent customer churn by using customer churn analysis PPT infographics. You can easily highlight information about the various marketing campaigns in order to retain its customer from churning. Provide ways to prevent churn through predictive analysis by incorporating our professionally designed customer churn prediction PPT presentation. https://bit.ly/3p6AR7S

RFM Analysis for Customer SegmentationCleverTap

RFM (Recency, Frequency, Monetary value) analysis measures a customer's recency of purchase, visit frequency, and typical monetary value. It helps marketers identify their best customers, customers at risk of churning, and potential customers. CleverTap's RFM dashboard visualizes customers on a 2D graph by recency and frequency scores. It segments customers into 10 groups including Champions, Potential Loyalists, New Customers, and At-Risk Customers to target personalized campaigns.

Customer Churn Management For Profit Maximization PowerPoint Presentation SlidesSlideTeam

The PowerPoint template allows firm in preventing its customers from reducing their purchase of products and services. It will help firm by providing various ways through which firm can manage their customer churn. It will cover details about churn propensity model. The template covers details about key statistics associated to customer churn. It covers details about present situation of customer attrition and customer churn rate on monthly basis. Customer churn is considered as critical issue which affect firms overall firm performance, due to which firm will incur heavy losses in terms of abandon purchases and lower revenues. It is to be noted that retaining customer is more profitable than acquiring new customers. The template will cover information regarding various types of customer churn such as when customer stops spending, churn due to product quality or complete customer account loss. The template will provide details about how firm will handle customer attrition by focusing on four stages of churn management by acquiring churned customers, delighting customers, preventing customer attrition, and saving customers through various campaigns. The template covers details about churn propensity model which will help preventing customer churn through predictive analytics by utilizing different statistical techniques such as machine learning. https://bit.ly/36qQZKg

RFM SegmentationKamil Bartocha

The document discusses RFM customer segmentation, which segments customers based on their Recency (time since last activity), Frequency (number of activities), and Monetary (total monetary value) metrics. These metrics can be calculated from transaction or engagement data and used to group customers into segments. The segments are identified by analyzing the distribution of the RFM metrics and identifying cut-off points. RFM segmentation allows identifying the most valuable customers and those at high risk of churn for targeted marketing campaigns.

Customer Life Cycle Model PowerPoint Presentation SlidesSlideTeam

The document provides an overview of a customer lifecycle model, including sections on customer segmentation, targeting audiences, understanding customer needs, customer acquisition strategies, marketing campaigns, customer retention, loyalty frameworks, satisfaction feedback, and key metrics. It includes examples of frameworks and processes for segmentation, targeting prospects, acquisition strategy, demand creation, sales force automation benefits, and the customer acquisition funnel. The slides are meant to be customized for a specific company.

Increasing the Lifetime Value of a CustomerAlan D Campbell

This document discusses strategies for increasing the lifetime value of customers. It begins by explaining how to calculate the annual profitability and lifetime value of customers. Customers are assets that provide future cash flows, so acquiring and retaining customers is important. The document then outlines 11 strategies for increasing a customer's lifetime value, such as increasing purchase frequency, sales amounts, customer retention, and developing new products/services that appeal to existing customers.

Churn prediction data modelingPierre Gutierrez

This document discusses how to model customer churn through machine learning. It defines churn as customers leaving or stopping usage. There are two types of churn - for subscription models where leaving can be clearly defined, and non-subscription models where leaving must be approximated. The document recommends predicting churn through classification models to identify potential churners, using customer behavioral and profile features over time. It also discusses evaluating models on validation data and using models to predict future churn and inform retention offers.

Data mining and analysis of customer churn datasetRohan Choksi

The document discusses a study conducted by a mobile phone company to analyze factors related to customer churn. The company provided a dataset of 3,332 customer records to build a neural network model that can predict which customers are likely to switch providers. Examining the data showed that increased usage of night, evening, and day minutes, as well as more customer service calls, correlated with higher churn. International calling plans also had a major impact on churn rates. The model achieved a misclassification rate of 7.11% and identified key variables for the company to address to reduce churn, such as international call pricing and infrastructure issues.

Rfm analysischintan desai

The document discusses the RFM model for customer behavior analytics. RFM stands for Recency, Frequency, and Monetary value. It is a technique used to segment customers based on how recently they made a purchase, how often they make purchases, and the total monetary amount of their purchases. The document provides details on calculating RFM scores, describes different customer life stages based on RFM, and highlights strengths of using RFM analysis, such as targeting the most profitable customers.

A Better Approach to Customer RetentionFramed Data

Customer Churn Analysis and PredictionSOUMIT KAR

This document discusses predicting customer churn in the telecom industry. It outlines collecting customer data from Kaggle on services used, account details, and demographics. Exploratory data analysis finds recent and higher spending customers more likely to churn. Feature selection and encoding is done before imbalanced data handling with SMOTE tomek. Various classifiers are tested with random forest performing best with an AUC of 0.834. Partial dependence plots and SHAP values are used to explain the model. Finally, a web app is created and deployed on Github to predict churn probabilities and help telecom companies reduce customer churn.

Ways to Reduce the Customer Churn RateFORMCEPT

This document discusses customer churn, which refers to the rate at which customers leave a company. It states that reducing churn by 5% can boost profits by 75% and that US companies lose $1.6 trillion per year to churn. Common causes of churn include poor customer service, price, functionality, and changing customer needs. The document promotes MECBot, a data analytics product that can detect, prevent, and reduce churn in real-time through capabilities like churn scoring, campaign management, and customer lifetime value maximization.

Churn and Customer LIfetime ValueJ. Scott Christianson

The document discusses churn, which is a measure of customer retention and loyalty. It defines churn as the number of customers lost over a period divided by the total number of customers at the beginning of the period. Churn is used to calculate key metrics like customer lifetime value (CLV) and is an important factor for subscription-based businesses. The document provides formulas for calculating churn rate, customer retention rate, customer lifetime period, and CLV. It also discusses how churn impacts a company's growth potential and provides tips for reducing different types of churn.

Telecommunication Analysis (3 use-cases) with IBM watson analyticssheetal sharma

1. The telecommunications company is concerned about customer churn and needs to understand the factors influencing why customers are leaving. 2. The analysis found that the two year contract has very few current customers and is influencing churn. Focusing on modifying the two year contract and its rate plans could increase customers. 3. For cross-selling and up-selling, the analysis found that prepaid customers are not using services broadly. Payment methods should be focused on and policies added to promote the company, such as discounts for cash payments or collaborating with banks.

Customer SegmentationIQBusiness_CEM

This document discusses customer segmentation and describes a customer needs-led segmentation approach. Traditional segmentation divides customers based on demographics but does not account for customer needs. The discussed approach identifies customer types based on their needs and priorities, defines successful customer outcomes, understands the customer lifecycle and preferred channels of engagement to develop new customer segments. Benefits of this approach include better customer insights, an aligned organization, efficient delivery, improved experience and increased revenue.

Customer AnalyticsBRIDGEi2i Analytics Solutions

Survival_AnalysisRushil Goyal

This document discusses using survival analysis to model customer churn. It explains what customer lifetime value is and why it is important. It then covers how survival analysis can be used to predict when customers will churn rather than just whether they will churn. The document outlines the methodology used, including data cleaning, defining churn, and graphing results. It also covers parametric survival models and evaluating different survival distributions. The conclusions state that survival analysis provides powerful insights into how attrition risk changes over time.

Customer Lifetime Value in Digital MarketingTaste Medio

This document discusses customer lifetime value (CLV) in digital marketing. It defines CLV as the present value of the future cash flows from a customer. For e-commerce companies, predicting customer behavior and future profit is more difficult than for companies in industries like telecommunications that have more predictable customer cash flows. The document provides examples of how CLV can be used to focus marketing efforts on profitable customers, evaluate direct marketing campaigns, and integrate customer data from CRM systems into digital advertising platforms like Google Ads to optimize bidding and targeting.

More Related Content

What's hot (20)

Customer Lifetime ValueJY Chun

Presentation on Customer Lifetime ValueKisorkumar Varadarajan

Customer lifetime valueyalisassoon

Churn Prediction in PracticeBigData Republic

Customer Churn Prevention Powerpoint Presentation SlidesSlideTeam

RFM Analysis for Customer SegmentationCleverTap

Customer Churn Management For Profit Maximization PowerPoint Presentation SlidesSlideTeam

RFM SegmentationKamil Bartocha

Customer Life Cycle Model PowerPoint Presentation SlidesSlideTeam

Increasing the Lifetime Value of a CustomerAlan D Campbell

Churn prediction data modelingPierre Gutierrez

Data mining and analysis of customer churn datasetRohan Choksi

Rfm analysischintan desai

A Better Approach to Customer RetentionFramed Data

Customer Churn Analysis and PredictionSOUMIT KAR

Ways to Reduce the Customer Churn RateFORMCEPT

Churn and Customer LIfetime ValueJ. Scott Christianson

Telecommunication Analysis (3 use-cases) with IBM watson analyticssheetal sharma

Customer SegmentationIQBusiness_CEM

Customer AnalyticsBRIDGEi2i Analytics Solutions

Customer Lifetime ValueJY Chun

Presentation on Customer Lifetime ValueKisorkumar Varadarajan

Customer lifetime valueyalisassoon

Churn Prediction in PracticeBigData Republic

Customer Churn Prevention Powerpoint Presentation SlidesSlideTeam

RFM Analysis for Customer SegmentationCleverTap

Customer Churn Management For Profit Maximization PowerPoint Presentation SlidesSlideTeam

RFM SegmentationKamil Bartocha

Customer Life Cycle Model PowerPoint Presentation SlidesSlideTeam

Increasing the Lifetime Value of a CustomerAlan D Campbell

Churn prediction data modelingPierre Gutierrez

Data mining and analysis of customer churn datasetRohan Choksi

Rfm analysischintan desai

A Better Approach to Customer RetentionFramed Data

Customer Churn Analysis and PredictionSOUMIT KAR

Ways to Reduce the Customer Churn RateFORMCEPT

Churn and Customer LIfetime ValueJ. Scott Christianson

Telecommunication Analysis (3 use-cases) with IBM watson analyticssheetal sharma

Customer SegmentationIQBusiness_CEM

Customer AnalyticsBRIDGEi2i Analytics Solutions

Similar to Predict Customer Lifetime Value Presentation (20)

Survival_AnalysisRushil Goyal

Customer Lifetime Value in Digital MarketingTaste Medio

Three case studies deploying cluster analysisGreg Makowski

Three case studies are discussed, that include cluster analysis as a component. 1) Customer description for a credit card attrition model, to describe how to talk to customers. 2) Hotel price optimization. Use clusters to find subsets of similar behavior, and optimize prices within each cluster. Use a neural net as the objective function. 3) Retail supply chain, planning replenishment using 52 week demand curves using thousands of seasonal "profiles" or clusters.

IRJET- Finding Optimal Skyline Product Combinations Under Price PromotionIRJET Journal

This document discusses finding optimal product combinations under price promotions. It introduces the concept of finding the k-most promising product combinations (k-MPP) from a set of products that will attract the maximum number of customers. An algorithmic framework is proposed to handle the k-MPP query and its variations. The framework uses lattice-based data partitioning and parallel computing techniques. Extensive experiments on real and synthetic datasets demonstrate the effectiveness and efficiency of the proposed framework.

Master defence 2020 - Anastasiia Kasprova - Customer Lifetime Value for Retai...Lviv Data Science Summer School

This document summarizes the author's master's thesis on developing models to calculate customer lifetime value (CLV) for a retail business based on transactional and loyalty card data. The thesis proposes both probabilistic and econometric CLV modeling approaches, applies clustering techniques like K-Means and Gaussian mixture modeling to segment customers, and uses Markov chains, time series analysis and survival models to estimate CLV and predict future business value. The frameworks are developed and tested on transaction data from 12 grocery stores over 1.5 years but have limitations from the short data time period. The thesis concludes by prototyping an analytical framework for offline retailers to estimate CLV from their operational data and use it for marketing evaluations.

Customer Segmentation Using Portfolio Optimization for B2B MarketsMd Mazedul Islam Khan

Mathematical Model For Customer Life Time Based Offer ManagementZehra Kendir

The document proposes a mathematical model for mobile operators to manage offers to prepaid customers using expected future transactions and customer lifetime value calculated from refill transaction data with a predictive model based on RFM analysis and the Pareto/NBD model. Several logistic regression models are used as benchmarks to evaluate the performance of the RFM data and additional customer features in predicting customer behavior. The results show that combining RFM data with additional customer features and customer segment and tenure information provides the most accurate predictions.

Юнит-Экономика CLV (customer lifetime value) #GFactors

CLV (Customer Lifetime Value) refers to the present value of all future profits obtained from a customer over their lifetime relationship with a firm. There are different approaches to defining and calculating CLV, as it can incorporate revenues, costs, relationship duration, and discount rates. Effectively using the CLV metric requires understanding concepts from marketing, economics, finance, and investment analysis. While CLV can provide useful insights, its application in practice faces challenges due to varying definitions and a lack of standardized methodology.

How to Ace Your Marketplace FundraiseSpeedinvest

Calculating Your Customer Lifetime Value - Dawn of the Data Age Lecture SeriesLuciano Pesci, PhD

Not all customers are created equal. Finding the average customer lifetime value is a quick win for your organization, and you definitely have the data necessary to do it, but a closer inspection of your data will show that 1 in 5 of your customers is responsible for 80% of the total lifetime value. And that isn't even taking non-monetary CLV into consideration. To beat your competition for the best customers you need individual-level insights about the value of each customer. This Lecture Will: -TEACH YOU HOW TO USE YOUR DATA TO FIND AVERAGE CLV. -EXPLAIN THE BEST PRACTICES WHEN DEALING WITH NON-MONETARY CLV. -SHOW YOU HOW TO MOVE BEYOND AVERAGE CLV TO INDIVIDUALIZED CLV. You can watch this lecture here: https://youtu.be/iCX-afWhmZ4

Data Con LA 2019 - Know Your Customer by Fiona LiData Con LA

In marketing, customer lifetime value (LTV, sometimes called CLV or CLTV) is a prediction of the net cash flow attributed to the entire future relationship with a customer. The prediction model can have varying levels of sophistication and accuracy, ranging from heuristic to the use of complex machine learning techniques. LTV in a non-contractual setting is widely accepted to be more difficult than in a contractual setting, in which the churn rate can be simplified as a constant. - As the world's largest concert promoter, Ticketmaster is focused on connecting live events to millions of our fans in a non-contractual and discrete setting. We adopted a paradigm called RFM (Recency, Frequency, Monetary) to make predictions of fans' two-year lifetime value to help make decisions including, for example - new product feature launch, SEM bidding automation, overall budgeting and essentially implement a winning strategy driven by customer lifetime value.- In this talk we will discuss RFM and other probabilistic models we used and how the results of the analysis helped drive business decisions. We will provide an overview of: (1) Customer lifetime value using RFM (2) Probabilistic Models: Bayesian Models, Beta-Geometric/Beta-Binomial Model (BG/BB) (3) A case study: how to use LTV for a new feature launch

How to-analyze-saas-businesses-gianluca-valentiniGianluca Valentini

Lean canvas @ Université TOTAL 2016Santiago LEFEBVRE

Customer Analytics Best PracticeTa-Wei (David) Huang

AUTM The Startup MetricsKeith McGreggor

This document discusses key startup metrics for evaluating the performance and health of a startup business. It introduces metrics related to customer acquisition cost (CAC), lifetime value (LTV), average revenue per user (ARPU), gross margin, churn rate, burn rate, and how these metrics can be used to determine if a startup will be profitable with a CAC lower than the LTV. Formulas are provided for calculating various metrics like CAC, LTV, and examples are given to demonstrate how metrics can indicate whether a startup business model is viable or not.

PredictingRepeatCustJawed Khan

This document describes using a decision tree model to predict repeat customers for an automotive company. It analyzes customer purchase history data from 2005-present to identify 125,629 repeat customers out of 425,745 total customers. A time window of 8 years is identified as having the highest likelihood of a second purchase. The decision tree model is trained on customer profile data and purchase history to assign probabilities of being a repeat customer. Highest probability customers within 8 years are prioritized for targeted solicitation through an automated dashboard process to improve sales.

Developing an effective LTV model at the soft launch and keeping it valid fur...GameCamp

Whole way of developing and maintaining an LTV model for Crazy Panda game starting from the very rough extrapolation models at the soft launch to more accurate user-based Machine Learning models for mature products. Moreover, we will peek into the main obstacles on our way and how to overcome them. How is LTV calculation different for new games at soft launch phase vs mature products? - Presentation run during on of GameCamp webinars; http://www.gamecamp.io/events/understanding-prediction-ltv/ - All GameCamp webinars: http://www.gamecamp.io/events/

Evolving Business Models in Digital HealthRobert Mittendorff, MD, MBA

Chapter 12: Applications of Database Marketing in B-to-C and B-to-B Scenariositsvineeth209

The document summarizes three studies on customer lifetime value and profitability relationships. Study 1 finds the relationship between lifetime and profits is not always positive, and both short-term and long-term customers can be profitable. Study 2 develops a model to estimate individual customer profitable lifetime durations based on projected revenues and costs. Study 3 analyzes factors that impact profitable lifetime durations, finding purchase amount, cross-buying, and mailings increase duration while returns and buying in a single department decrease it.

Capstone_Customer_Churn_Prediction_Capstone_proejct.pptxBalkrishnaVasudev

Survival_AnalysisRushil Goyal

Customer Lifetime Value in Digital MarketingTaste Medio

Three case studies deploying cluster analysisGreg Makowski

IRJET- Finding Optimal Skyline Product Combinations Under Price PromotionIRJET Journal

Master defence 2020 - Anastasiia Kasprova - Customer Lifetime Value for Retai...Lviv Data Science Summer School

Customer Segmentation Using Portfolio Optimization for B2B MarketsMd Mazedul Islam Khan

Mathematical Model For Customer Life Time Based Offer ManagementZehra Kendir

Юнит-Экономика CLV (customer lifetime value) #GFactors

How to Ace Your Marketplace FundraiseSpeedinvest

Calculating Your Customer Lifetime Value - Dawn of the Data Age Lecture SeriesLuciano Pesci, PhD

Data Con LA 2019 - Know Your Customer by Fiona LiData Con LA

How to-analyze-saas-businesses-gianluca-valentiniGianluca Valentini

Lean canvas @ Université TOTAL 2016Santiago LEFEBVRE

Customer Analytics Best PracticeTa-Wei (David) Huang

AUTM The Startup MetricsKeith McGreggor

PredictingRepeatCustJawed Khan

Developing an effective LTV model at the soft launch and keeping it valid fur...GameCamp

Evolving Business Models in Digital HealthRobert Mittendorff, MD, MBA

Chapter 12: Applications of Database Marketing in B-to-C and B-to-B Scenariositsvineeth209

Capstone_Customer_Churn_Prediction_Capstone_proejct.pptxBalkrishnaVasudev

Recently uploaded (20)

Thingyan is now a global treasure! See how people around the world are search...Pixellion

Adobe Analytics NOAM Central User Group April 2025 Agent AI: Uncovering the S...gmuir1066

Molecular methods diagnostic and monitoring of infection - Repaired.pptx7tzn7x5kky

GenAI for Quant Analytics: survey-analytics.aiInspirient

md-presentHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHation.pptxfatimalazaar2004

Ppt. Nikhil.pptxnshwuudgcudisisshvehsjkspanchariyasahil

Developing Security Orchestration, Automation, and Response ApplicationsVICTOR MAESTRE RAMIREZ

1. Briefing Session_SEED with Hon. Governor Assam - 27.10.pdfSimran112433

chapter 4 Variability statistical research .pptxjustinebandajbn

Data Analytics Overview and its applicationsJanmejayaMishra7

Stack_and_Queue_Presentation_Final (1).pptxbinduraniha86

Safety Innovation in Mt. Vernon A Westchester County Model for New Rochelle a...James Francis Paradigm Asset Management

By James Francis, CEO of Paradigm Asset Management In the landscape of urban safety innovation, Mt. Vernon is emerging as a compelling case study for neighboring Westchester County cities. The municipality’s recently launched Public Safety Camera Program not only represents a significant advancement in community protection but also offers valuable insights for New Rochelle and White Plains as they consider their own safety infrastructure enhancements.

Flip flop presenation-Presented By Mubahir khan.pptxmubashirkhan45461

Perencanaan Pengendalian-Proyek-Konstruksi-MS-PROJECT.pptxPareaRusan

04302025_CCC TUG_DataVista: The Design Storyccctableauusergroup

Simple_AI_Explanation_English somplr.pptxssuser2aa19f

DPR_Expert_Recruitment_notice_Revised.pdfinmishra17121973

Data Science Courses in India iim skillsdharnathakur29

This comprehensive Data Science course is designed to equip learners with the essential skills and knowledge required to analyze, interpret, and visualize complex data. Covering both theoretical concepts and practical applications, the course introduces tools and techniques used in the data science field, such as Python programming, data wrangling, statistical analysis, machine learning, and data visualization.

IAS-slides2-ia-aaaaaaaaaaain-business.pdfmcgardenlevi9

How to join illuminati Agent in uganda call+256776963507/0741506136illuminati Agent uganda call+256776963507/0741506136

Thingyan is now a global treasure! See how people around the world are search...Pixellion

Adobe Analytics NOAM Central User Group April 2025 Agent AI: Uncovering the S...gmuir1066

Molecular methods diagnostic and monitoring of infection - Repaired.pptx7tzn7x5kky

GenAI for Quant Analytics: survey-analytics.aiInspirient

md-presentHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHation.pptxfatimalazaar2004

Ppt. Nikhil.pptxnshwuudgcudisisshvehsjkspanchariyasahil

Developing Security Orchestration, Automation, and Response ApplicationsVICTOR MAESTRE RAMIREZ

1. Briefing Session_SEED with Hon. Governor Assam - 27.10.pdfSimran112433

chapter 4 Variability statistical research .pptxjustinebandajbn

Data Analytics Overview and its applicationsJanmejayaMishra7

Stack_and_Queue_Presentation_Final (1).pptxbinduraniha86

Safety Innovation in Mt. Vernon A Westchester County Model for New Rochelle a...James Francis Paradigm Asset Management

Flip flop presenation-Presented By Mubahir khan.pptxmubashirkhan45461

Perencanaan Pengendalian-Proyek-Konstruksi-MS-PROJECT.pptxPareaRusan

04302025_CCC TUG_DataVista: The Design Storyccctableauusergroup

Simple_AI_Explanation_English somplr.pptxssuser2aa19f

DPR_Expert_Recruitment_notice_Revised.pdfinmishra17121973

Data Science Courses in India iim skillsdharnathakur29

IAS-slides2-ia-aaaaaaaaaaain-business.pdfmcgardenlevi9

How to join illuminati Agent in uganda call+256776963507/0741506136illuminati Agent uganda call+256776963507/0741506136

Predict Customer Lifetime Value Presentation

1. A BIRD’S EYE VIEW OF CUSTOMER LIFETIME VALUE

2. About me

3. Contents 1. Why? 2. RFM & BG/NBD (2005) 3. Groupon: Churn + Random Forest (2016) 4. Asos: Random Forest + Embeddings (2017)

4. Customer Lifetime Value Customer Lifetime Value (CLTV/CLV/LTV) is the net present value of the future (net) cash flows associated with a particular customer. Peter Fader, Customer Centricity, p 72, Wharton Digital Press, 2012

5. Why Use Customer LTV? ● It’s not (always) an aggregate metric ● Not all customers are equally profitable ● Cost of acquisition can be 5x cost of retention ● What about the time horizon? ○ 1 month future LTV? ○ 3 month future LTV? ○ 1 year future LTV? ○ 1 decade future LTV? ○ 5 decades future LTV?

6. Business Context Contractual Non-Contractual ● Customer “death” can be observed ● Often modeled using survival-based approaches ● Customer “death” is unobserved ● Customer lifetime distribution often modeled via exponential models ● Amazon Prime ● Spotify Premium ● Netflix ● Gym membership ● Health Insurance ● Playstation+ ● Buying stuff from Amazon ● Buying stuff on Ebay ● Buying stuff from Tesco ● Buying video games on Steam ● Buying top ups on Ding

7. R F M LTV 5 5 4 2 2 3 1 2 5 ● Recency ○ How long ago did the customer last transact? ● Frequency ○ How many times have they transacted? ● Monetary Value ○ How much revenue did they generate? Pareto (BG)/NBD Fader et al, 2005

8. Pareto BG/NBD

9. Pareto BG/NBD Assumptions ● P(Alive | RFM, Timeframe) ● E[Transactions | RFM, Timeframe]

10. Pareto BG/NBD - Decent benchmark - Not appropriate for contractual settings - Not appropriate when transactions can only incur at fixed intervals (eg festivals)

11. 80/20 rule in action

12. Pareto Distribution https://www.kaggle.com/c/ga-customer-revenue-prediction/

13. Groupon - 2 stage random forest: Churn + CLV - 40 features including where ɑ and β are historical conversion rates for opened and clicked emails - User segments (never purchased | new | one-time | sporadic users | power users) Vanderveld et al, 2016

14. Groupon -- Model Performance

15. ASOS Chamberlain et al, Customer Lifetime Prediction Using Embeddings, KDD, 2017 AUC Churn: 0.79 Spearman CLV: 0.56

16. ASOS -- Cost Maximum Area Under the receiver operating characteristics Curve (AUC) achieved on a test set of 50,000 customers in hybrid models against the number of neurons in the hidden layers (in log scale). Mean monetary cost to train hybrid models on a training set of 100,000 customers against the number of neurons in the hidden layers (both in log scale).

17. Conclusions ● Revenue distribution is not particularly friendly for making predictions ● RFM & BG/NBD is a good start & benchmark ● Random Forest models probably deliver the most bang for the buck ● Neural Nets can improve predictions, but at a significant cost

18. References ● Peter Fader talk on Pareto BG/NBD ● Counting Your Customers the Easy Way [pdf] ● Python Lifetimes Library ● Etsy CLV [pdf] ● 2006 review by Sunil Gupta et al ● An Engagement-Based Customer Lifetime Value System for E-commerce [pdf] ● Customer Lifetime Value Prediction Using Embeddings [pdf] ● Implementing and Training Predictive Customer Lifetime Value Models in Python ● Write-up of the link above

19. Thank you Eric Mehes @meHessian https://www.linkedin.com/in/eric-mehes/

Predict Customer Lifetime Value Presentation

Recommended

More Related Content

What's hot (20)

Similar to Predict Customer Lifetime Value Presentation (20)

Recently uploaded (20)

Predict Customer Lifetime Value Presentation