Customer Analytics Case Study Report

DATA4700

Digital Marketing and

Competitive Advantage

Lesson 3

Customer Relationship Management

1 Measure customer churn and understand the

reasons why it occurs

2 Be familiar with analytics techniques to predict

customer churn

3 Implement interventions to prevent customer

churn from happening

4 Understand the relationship between customer

churn and customer relationship management

Lesson Learning Outcomes

Review of Lesson 2 1. Positioning (from the STP framework)

– Establishing a unique image:

• in the mind of your target customers

• for your company and your products

• that is completely different to the image of your competitors

2. Key framework #3: The 4 Ps of the Marketing Mix support the positioning

– Product: What features should the product have?

– Promotion: Where do we tell target consumers about the product?

– Place: How do we get the product to these consumers?

– Price: How much do we sell the product for?

3. The Digital Marketing Portfolio is used for promotion of your company and products

– Devices

• The physical tools used for engagement

– Platforms

• The host sites on which companies interact with customers

– Media

• The format through which the message is transmitted

• Owned, paid and earned media

3

Key framework #1: The 3 Cs – customers, company and competitors

Key framework #2: STP – segmentation, targeting, positioning

Customer churn

Customer churn is a metric of how many customers stop

using your product or service.

• It acts as a useful benchmark to determine whether your

marketing activities are resulting in desired outcomes

• Also known by the terms “customer attrition”, “customer

turnover” and “customer defection” (especially if you are

losing customers to a competitor)

Calculating churn rates

To calculate customer churn rates:

Step 1:

Choose whether you wish to measure number of customers,

sales revenue, or another important business metric (such as

sales profit)

Step 2

Choose your timeframe – this could be measuring:

• current month versus a previous month

• current quarter versus a previous quarter

• current year versus previous year

Example: Calculating churn rates

Customer churn is calculated according to the following formula:

Number of customers lost

Number of customers at start of period

Let’s say that you own an on-demand video subscription service

that features documentaries made by independent filmmakers.

https://variety.com/2018/tv/news/nab-local-tv-broadcast-fake-news-survey-1202747108/

https://www.acsh.org/news/2017/08/18/netflix-last-place-youll-find-pro-science-documentary-11697

Example: Calculating churn rates

Number of customers lost

Number of customers at start of time period

On 31 March, you saw in your database that you had a

total of 4,726 subscribers to your service.

On 30 April, your database revealed that the number of

subscribers had decreased to 4,489.

What is your customer churn rate?

Number of customers lost

Number of customers at start of time period

On 31 March: 4,726 customers

On 30 April: 4,489 customers

Number of customers lost: 4,726 − 4,489 = 237

Number of customers at start of time period (31 March): 4,726

Customer churn rate: 237 divided by 4,726

237 = 0.05 ( 5%)

4,726

Example: Calculating churn rates

The table below shows data from your video subscription service database over

the past six months:

1. Calculate the customer churn rate from end of September to end of October.

2. What other churn rates can you calculate?

Compare your answer for Q1 with a classmate, and then discuss possible

answers for Q2.

Be prepared to share your insights with the rest of the class.

Breakout activity: Calculating churn rates

Month ending No. of subscribers Revenue received

31 May 4,392 $193,248

30 June 4,306 $185,159

31 July 4,187 $177,947

31 August 4,101 $172,160

30 September 3,984 $166,332

31 October 3,905 $161,863

Class discussion: Calculating churn rates

Reminder: write in your worksheets!

1. Customer churn rate from end of September to end of October:

2. Other churn rates that you calculated:

Remember

Step 1: Choose whether you wish to measure number of customers,

sales revenue, or another important business metric.

Step 2: Choose your timeframe.

Customer churn –

why should we care?

Why is customer churn important?

• Reason #1: Difference in revenue generated over time.

The chart below shows the revenue of a company making 0.25%

revenue each month (blue line), compared to what its revenue

would be if it suffered a 5% churn each month (green line).

Source: https://medium.com/

@brooke.land/the-true-cost-of-

churn-3470359cd09

Customer churn –

why should we care?

Why is customer churn important?

• Reason #2: It often costs much

more to recruit new customers

than it does to retain existing

customers.

Sources:

• https://www.forbes.com/sites/jiawertz/2018/09/12/don

t-spend-5-times-more-attracting-new-customers-

nurture-the-existing-ones/#105950015a8e

• https://www.invespcro.com/blog/customer-acquisition-

retention/

Customer churn –

why should we care?

Why is customer churn important?

• Reason #3: Recurring customers tend to spend more.

Source: https://media.bain.com/Images/Value_online_customer_loyalty_you_capture.pdf

Breakout activity: Customer

churn – why should we care?

Watch the following video about market share and competitive advantage.

Video: Investopedia: How Can Companies Increase Market Share?

https://www.investopedia.com/ask/answers/031815/what-strategies-do-companies-

employ-increase-market-share.asp

1. How does customer churn relate to competitive advantage?

2. What benefits do companies get from this competitive advantage?

3. How can you cause your competitors to suffer customer churn?

In other words, how can you attract customers away from your

competitors and towards your company?

Discuss in groups of two or three, and be prepared to

share your insights with the rest of the class.

Class discussion: Customer

churn – why should we care?

Reminder: write in your worksheets!

1. Relationship between customer churn and competitive advantage:

2. Benefits from having a competitive advantage:

3. How to cause your competitors to suffer customer churn:

Using analytics to predict

customer churn

In order for analytics techniques to produce useful insights, companies

must collect pertinent data about their customers, which includes:

• how customers use their products, in current and previous periods

• records from sales and customer service contacts

• what customers are saying about the company and its products, either

publicly through social media or to the company directly (through online

surveys and other customer feedback channels)

What other relevant data might

companies want to collect?

https://www.hgsdigital.com/blogs/customer-data-platform/

Using analytics to predict

customer churn

What analytics techniques are available to predict customer churn?

Machine learning is a popular technique in the area of predictive

analytics to anticipate customer churn.

If the business goal is to recognise subscribing patterns in the customer

data:

• companies have data about which customers renew their

subscriptions (these are retained customers who provide recurring

revenue)

• companies have data about which customers do not renew their

subscriptions (these people form your customer churn)

• algorithms use the data to predict which current customers, based on

their characteristics, will likely stop being future subscribers

Three main areas of data analytics

• Descriptive – summarizing what has happened in the past

• Predictive – anticipating what will happen in the future

• Prescriptive – determining what should happen for an optimal outcome

Machine learning is associated with predictive analytics

Using analytics to predict

customer churn

What analytics techniques are available to predict customer churn?

There are two main machine learning techniques we use to predict

customer churn:

1. Decision tree, which seeks to categorise a current customer as

someone who will continue or not continue subscribing at a future time.

Output: Customer #18752 is classified as a non-subscriber next

month.

2. Logistic regression, which seeks to assign a likelihood that a current

customer will not continue subscribing at a future time.

Output: Customer #18752 has a 54% probability of not continuing the

subscription next month.

Using analytics to predict

customer churn

What analytics techniques are available to predict customer churn?

1. Decision tree, which seeks to categorise someone’s actions in the future.

Example: Classifying someone as passing or failing tomorrow’s exam.

https://www.mihaileric.com/posts/decision-trees/

Using analytics to predict

customer churn

What analytics techniques are available to predict customer churn?

2. Logistic regression, which seeks to assign a likelihood.

Example: Predicting the probability of someone passing tomorrow’s exam.

http://www.presentica.com/ppt/115434/logistic-regression-279676

Breakout activity: Using analytics

to predict customer churn

Watch the following short video.

Video: Microsoft: Customer Churn Prediction

https://www.youtube.com/watch?v=xTWnoAwyHPs

1. What data about the customers of Tom’s Pet Store could be input into a

model to predict customer churn?

2. Thinking about the on-demand video subscription service whose churn

you calculated earlier, what data about your customers could be input

into a model to predict your customer churn?

Discuss in groups of two or three, and be prepared to share your insights

with the rest of the class.

Class discussion: Using analytics

to predict customer churn

Reminder: write in your worksheets!

1. Data about Tom’s customers that could be input into a predictive

model:

2. Data about the customers of your subscription service that could

be input into a predictive model:

Preventing customer churn

Customers leave for a reason; if you can identify and understand

what that reason is, you are more likely to prevent that customer

from churning.

Common reasons for churning:

• Bad user experience with the product

• Bad customer service experiences

What do customers do after they have a bad user or customer

service experience?

• Bad fit: the product no longer addresses

the customer’s needs (or never did)

https://plainmagazine.com/15-useless-product-designs/

Preventing customer churn

How quickly do customers respond to bad experiences?

https://www.pwc.com/us/en/advisory-services/publications/consumer-intelligence-series/

pwc-consumer-intelligence-series-customer-experience.pdf#page=8

Preventing customer churn

How important is customer experience to purchasing decisions?

https://www.pwc.com/us/en/advisory-services/publications/consumer-intelligence-series/

pwc-consumer-intelligence-series-customer-experience.pdf#page=8

Preventing customer churn

What can you do to stop customer churn?

1. Give customers plenty of opportunities

to provide feedback:

– Listen to your customers, especially at key stages in the

customer lifecycle (e.g. shortly after first purchase; when usage

starts to drop)

2. Be proactive about communication and customer service

– Anticipate problems before they

arise for customers

– Don’t take customers for granted

3. Monitor the Net Promoter Score

(NPS)

https://www.qminder.com/ask-customers-feedback/

https://www.researchgate.net/figure/Calculating-the-Net-Promoter-Score_fig1_326787367

Preventing customer churn

What can you do to stop customer churn?

4. Define a roadmap for your customers

– Onboarding for new customers

– Higher levels of recognition for deeper engagement

5. Make your customers realise the value of your offering

– Emphasise the benefits rather than the features

6. Manage customers’ expectations

– Set expectations early and often

– At least meet those expectations; exceed them where possible

– Avoid over-promising if you can’t deliver

https://www.executivetraveller.com/bankwest-cuts-credit-card-qantas-frequent-flyer-points

Breakout activity:

Preventing customer churn

Watch this short video about the Starbucks Reward program.

Video: My Starbucks Rewards - Now on Android and iOS

https://youtu.be/lp4FRAkHMbc

1. What are some key features of the Starbucks Rewards program?

2. What customer data does Starbucks collect through this program?

Discuss in groups of two or three, and be prepared to

share your insights with the rest of the class.

https://www.starbucks.com/rewards/

Class discussion:

Preventing customer churn

Reminder: write in your worksheets!

1. Key features of the Starbucks Rewards Program:

2. Customer data that Starbucks collects through this program:

Customer relationship management

Customer relationship management (CRM) systems serve as a

central hub to help you:

• predict which customers are at risk of churning

• act to prevent those customers from churning

• manage customer loyalty to reduce churn rate

http://www.xobber.com/evolution-phases-of-crm-system/

Customer relationship management

Thinking strategically about CRM:

Source: Adrian Payne and Pennie Frow (2005). “A Strategic Framework for Customer Relationship

Management.” Journal of Marketing 69(4), 167-176.

www.jstor.org/stable/30166559

Customer relationship management

Good CRM systems do the following:

1. Generate insights about your customers, both on an individual

level as well as within the segments they belong to

2. Optimise interactions with customers at their preferred touchpoints

3. Provide reporting functionalities that help you understand the data

and help you persuade key stakeholders of your recommendations

4. Deliver the Marketing Mix appropriate for each customer:

– product recommendations

– relevant promotions

– suitable pricing

– where to access the product

Customer relationship management

Workforce benefits to the company of using a CRM system:

1. Sharing data across the organisation

– Works to break down the silo effect

– Improves teamwork and collaboration

2. Increased visibility and accountability

– All workers own their responsibilities

– Who is doing what? Who should be doing what?

3. Automation of more operational tasks

– Frees up time for workers to engage in human interactions with

customers

https://hrmasia.com/hrm-five-elements-of-a-high-performance-team/

Widely used CRM systems

SAS Customer Intelligence 360 Salesforce

Microsoft Dynamics Zoho

https://www.sas.com/en_au/software/customer-intelligence-360.html

https://www.salesforce.com/au/crm/

https://dynamics.microsoft.com/en-us/ai/customer-insights/

https://www.zoho.com

Summary of Lesson 3 1. Customer churn

– Decreases revenue

– Expensive to recruit new customers

– Existing customers spend more with you

2. Predicting customer churn with machine learning

– Decision tree to classify

– Logistic regression to assign a probability

3. Preventing customer churn

– Customers leave due to bad experiences, customer service, or product fit

– Customer loyalty keeps them engaged

4. Customer relationship management (CRM)

– “Holistic approach to managing customer relationships to create shareholder value”

– Workforce benefits for the company as well

Key framework #1: The 3 Cs – customers, company and competitors

Key framework #2: STP – segmentation, targeting, positioning

Key framework #3: The 4 Ps – product, promotion, place, price

Digital marketing portfolio: devices, platforms, media format and type

Case study: Philips

Watch this short video about Philips, the Dutch health technology

and home electronics company.

Video: Salesforce: Philips is a Trailblazer

https://youtu.be/b6ntEatqj3k

1. What stands out to you about Philips’ business approach?

https://www.philips.com/global

Appendix: Customer churn

data workshop

Predicting Customer Churn

Using Spotfire to predict whether a customer (mobile services) will

leave the company.

We will:

• Explore the data to identify pertinent features.

• Perform logistic regression on a subset of variables.

• Develop a classification tree on a subset of variables.

• Analyse the outcomes and comment of the results.

• Examine implications for further studies.

Predicting Customer Churn

IMPORT

Load the file ‘cell2cell_churn’ into Spotfire.

Predicting Customer Churn IMPORT

The following is a summary of field descriptions for select fields. Field DataType Description

Churn Categorical Did the customer switch to a different provider?

MonthlyRevenue Numeric Average monthly spend on phone services.

MonthlyMinutes Numeric Average number of minutes of call per month.

DroppedCalls Numeric Average number of calls that lost connection per month.

CustomerCareCalls Numeric Average number of calls to customer service per month.

ActiveSubs Numeric Number of services the customer subscribes to.

Handsets Numeric Number of handsets the customer has had over time.

HandsetModels Numeric Number of different handset models the customer has had over time.

CurrentEquipmentDays Numeric Number of days the customer has had their current handset.

HandsetWebCapable Binary Is the handset able to browse the web.

OwnsComputer Binary Whether the customer owns a computer.

HasCrediCard Binary Whether the customer has a credit card.

Credit Rating Categorical A scale of 1 to 7, 1 being highest quality and 7 being lowest quality.

Income Group Categorical A scale of 1 to 9, 1 being the lowest income stratum and 9 being the highest.

Prism Code Categorical A reclassification of the user’s residential post code into one of Rural, Suburban,

Town or Other.

Predicting Customer Churn

EXPLORE

Create a stacked barchart to compare where customers live

(PrizmCode) and their socioeconomic status (IncomeGroup)…

Predicting Customer Churn

EXPLORE

Create a stacked barchart to compare creditworthiness (CreditRating)

and their socioeconomic status (IncomeGroup)…

Predicting Customer Churn

EXPLORE

Create a stacked barchart to compare credit card possession

(HasCreditCard) and their socioeconomic status (IncomeGroup)…

Predicting Customer Churn

EXPLORE

Heatmap – This produces a matrix that can show the average monthly

revenue for customers as defined by their credit rating and income

level.

Predicting Customer Churn

EXPLORE

Heatmap – We can also examine the topology for monthly revenue by

credit card possession and credit rating.

Predicting Customer Churn

EXPLORE

Waterfall Chart – What is the monthly revenue contribution by income

group?

Predicting Customer Churn

EXPLORE

Waterfall Chart – What is the monthly revenue contribution by credit

rating segments?

Predicting Customer Churn

EXPLORE

KPI Chart – average monthly revenue and handsets per customer by

income group.

Predicting Customer Churn

ANALYSE – LOGISTIC REGRESSION

Select ‘Classification modelling’ from the Tools menu.

Give your model a name

Add a short comment

Choose ‘Logistic Regression’

Select ‘Churn’ as the

response column

Choose what you believe are

good predictors of churn.

Use the ‘Add’ button and

select variables of interest

to add to the model.

Select OK when ready

Predicting Customer Churn ANALYSE – LOGISTIC REGRESSION

Akaike Information Criterion: A measure of overall model efficacy.

Predictor coefficient: The size of the response per unit change in the predicator.

P-value (significance): How significant is the independent (predictor) variable.

Akaike Information Criterion

Predictors of customer churn

Probability of significance

Predictor coefficients

Predicting Customer Churn Summary – LOGISTIC REGRESSION

Akaike Information Criterion:

• A measure of overall model efficacy.

• Used to compare against other models.

• The lower the better.

Predictor coefficient:

• The size of the response per unit change in the predicator.

• E.g.: ‘CreditRating5-Good’ has coefficient -0.16, suggesting customers with

good crediting ratings tend to leave this company.

• E.g.: ‘CreditRating2-Low’ has coefficient +0.25, suggesting customers with

substandard crediting ratings tend to stay with this company.

P-value (significance):

• How significant is the independent (predictor) variable.

• At 95% confidence, we look for p ≤ 0.05

Predicting Customer Churn

ANALYSE – CLASSIFICATION TREE

Select ‘Classification modelling’ from the Tools menu.

Give your model a name

Add a short comment

Choose ‘Classification

Tree’

Select ‘Churn’ as the

response column

Choose what you believe

are good predictors of

churn.

Use the ‘Add’ button and

select variables of interest

to add to the model.

Select OK when ready

Predicting Customer Churn ANALYSE – CLASSIFICATION TREE

Notable outputs are:

ROC curve – Indicative of model efficacy, ideally it should bend.

Predicted outcome probabilities – a probability for each response, sums to 1.

Note: classification trees do not produce AIC statistics so ROC becomes basis

for comparison.

Churn probabilities:

p(NO) = 71.2%

p(YES) = 28.8%

ROC curve of a good model

ROC curve of our

model indicates a

poor fit.