Sir Richard Branson has just welcomed you into his spasious office at Virigin Media headquarters in London. Quite concerned with many of his customers switching to competitors he said: „Recently we have found out that many customers have left us for our competitors. My business intelligence department is abroad on a team-building event and until they come back I am concerned that more customers will leave. This situation requires immediate, insight-driven actions that will stop and hopefully reverse customer churn. Therefore, I have a couple of analytical tasks that I need solved in the best of your capacity. Once you create insight, I expect that you come up with a set of strategic and tactical recommendations on how to act to prevent or possibly reverse customer churn.
Task 1: Identify who of our current customers is next to leave Virgin Media. You are provided with two files (database extracts). The first one VirginMediaChurn_Jan2018.csv includes information about:
Customers who left Virgin Media in January 2018 – the column is called ChurnServices and extras that each customer has signed up for – Landline, MultipleLines, Broadband, Powerline Kit, Online Backup, Gadget Rescue, Cable TV and Video On DemandCustomer account information – how long they have been a customer (tenure), contract type, payment method, paperless billing, monthly charges, and total chargesDemographic info about customers – gender, employment and marital status and if they have any dependents
The second file is VirginMedia_Feb2018.csv includes everything listed above except Churn. Building on the knolwedge obtained from the first file, try to predict who of our current customers is likely to leave. To build prediciton model, choose the best prediction technique, explain which variables are most important to predicting churn and establish who will churn. Provide a detaled explanation of your analytical approach – i.e. workflow model and its elements. Explain step by sep what was done to get the best possible predicitons. Provide detailed overview of the outputs and elaborate your findings. Provide links to your workflows and files (be explicit in which file you have saved predicted churners). Do not forget that my DBI team is very picky so make sure you do a rock solid modelling and analysis.
Task 2: Once you have established who is likely to churn, check if different types of churners exist. If so, establish how many segments there are, profile identified segments using available data and give appropriate names to the segments. Same as in the previous task, elaborate your approch and your findings. Provide link to your workflow model and files (be explicit in which file you have saved your classification results).
Task 3: I also want to know WHY our customers leave. Unfortunately, we cannot see this in our databases and we have no time to commission market research. Instead, I would like you to see what have people tweeted recently about VirginMedia. This might signpost to the potential churn causes. I would like to see key themes and sentiment of the tweet corpus mentioning our business name. Anything above this minimum requirement would be a big plus. Same as in the previous tasks, elaborate your approch and your findings. Provide links to your workflow models and files.
Once you have solved all the tasks, you need to provide recommendations on what actions VirginMedia should take to prevent and potentionally reverse churn. While you have creative freedom, your recommendations have to be linked with the insight you have created and other external research. Please note that there will be two audiences reading this report. First audience is made of myself and my board of directors. Given that we are extremely busy we do not like technical and leghty reports that require hours of reading. Instead we prefer intuitive infographs, with very little text that can be easily understood. We are willing to read up to 3 infographic slides. Please put these in the executive summary so that we do not have to struggle finding them in your report. The second audience is my DBI team. They will be interested to throroughly read and inspect your analytical approach and findings. Therefore, be elaborate and precise and provide links for my team to inspect your model workflows, working files and outputs. Justify your decisions in building models and demonstrate that you know how different widgets/tools function. And once again, bear in mind you do all of these analyses to help us stop the churn. Good luck!“
Sir Richard Branson, Virgin Media
Further instructions and expectation
All analyses must be performed in Orange Data Mining software. If necessary install add-ons that will assisst you in completeing your tasks. In explaining your analytical approach you are allowed to insert workflow screenshots. It is vital that your data mining models are well annotated and made available for your tutors to download. Therefore, save all your workflows and data files in cloud (Dropbox, Google drive etc.) and provide links to each task’s solution so that your tutors can inspect it when marking. When making infographics you can use any software you want. Please note, you are not allowed to write to or call Virgin Media or Sir Richard Branson in any way or form. Please note, producing digital business intelligence report is a very demanding task. Start working on it as soon as possible and make sure that you do not delegate tasks back to the client.
Suggested report structure
Digital Business Intelligence reports can take various forms and structures. We would recommend but not insist that your report follows some of the structures given below. You can make your report in MS Word or MS PowerPoint with later giving you more flexibility in building infographics. However, the final report structure and report organisation is ultimately your decision. The suggested plan structures:
Alternative 1:Alternative 2:Cover pageTable of contentsExecutive summary infographics1. Task 1 1.1. Approach 1.2. Findings2. Task 2 2.1. Approach 2.2. Findings3. Task 3 3.1. Approach 3.2. Findings4. RecommendationsReferencesAppendices Cover pageTable of contentsExecutive summary infographics1. Task 1 1.1. Approach 1.2. Findings 1.3. Recommendations2. Task 2 2.1. Approach 2.2. Findings 2.3. Recommendations3. Task 3 3.1. Approach 3.2. Findings 3.3. Recommendations4. ConclusionReferencesAppendices
For each task please provide a link to your workflows and working files. These can be stored on Dropbox, Google drive or any other cloud based drive/location. Once your report is submitted you are not allowed to alter your workflows and files. Dates on files will be checked. Screenshots of your workflows and your actual workflows must not differ. Otherwise your tasks will not be marked at all.
Teamwork instructions and group allocations
Groups can consist of a maximum of 3 students. It is your responsibility to find a teammate for digital business report within your seminar/workshop group. If you struggle to find a partner for your coursework, please email your seminar tutor immediately (make sure you include your details as well as your seminar group). Your module leader will, if possible, pair you with another student seeking a partner. Waiting for too long to find a coursework partner will put you in the risk of working on the plan on your own. Failure to inform the module leader of their desired pairing by the stated date will mean that students are allocated into a randomly selected group. If this happens there will be no further opportunity to switch groups.
Team members should contribute equally and meet on a regular basis to work together on business intelligence report. Make sure you keep your meeting minutes accurate and up to date. To do so, you can use the form provided below to ensure a good team management. In case of a team dispute, meeting minutes can be a good starting point for a resolution.
Group Presentation Mark Allocation
The module tutors will assess reports and award each group a mark out of 100. You will then each be asked to assess your group members’ and your own contribution and allocate marks accordingly. To do this, we will be using the University’s Web Peer Assessment (WebPA) system. WebPA is an online tool which enables students in a group to peer review. Group work is often criticised because each student receives the same team mark regardless of individual performance. By using WebPA, peers assess each member of their group’s contribution to their work and then each student receives an adjusted mark. If a student has worked hard, contributed consistently throughout the group project and been committed to the group, then they will receive a better mark than their peers who have not contributed to the same extent. For example, if your tutor grades a group with 66%. Normally all the members of the group would receive 66% regardless of how much they had contributed to the group’s work. With online peer assessment, a percentage of the mark is based on the tutor’s grade and the remainder on peer assessment. One student, Mary, who has worked hard and contributed to the progress of the group may have her grade increased to 73% through peer marking (if her peers acknowledge her commitment by ranking her highly on all the criterion) but someone who has not been so committed may find that her grade is reduced to 59%. Tutors can check and moderate peer scores and override if appropriate. In other words, every student in the group has an identical ‘cake’, but they can give different sized slices to each other. A full worked example of the algorithm explaining how these grades are worked out is available at:http://webpaproject.lboro.ac.uk/?q=node/125. All students will be asked to complete their peer reviews immediately after report submission. Failure to complete your peer review within the allocated timeframe will result in that individual student receiving a penalty of -10% of their mark. All students are also advised to complete the process individually and away from their peers to ensure they provide an assessment that truly reflects each member’s contribution. Please note that due to the nature of the system, students will only get one opportunity to complete their peer assessment form. There are no opportunities to reset this if you make a mistake so ensure that you fully concentrate when completing this assessment. A demonstration of the process will be provided within one of the lectures to assist with this process.
The post Understanding Virgin Media customer churn
Assignment status: Solved by our experts