Telecommunication Analysis (3 use-cases) with IBM watson analytics
2 likes2,961 views
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.
Telecommunication Analysis (3 use-cases) with IBM watson analytics
- 1. Contents Introduction Purpose Task 1 – Churn Analysis Task 2 – Cross-selling and Up-selling Task 3 - Customer Segmentation & Customer Lifetime Value Analytics
- 2. Introduction A telecommunications company is concerned about the number of customers leaving their Services. They need to understand who is leaving and why they leaving. The company's customer database includes customers who churned within the last month. In addition to information on Churn (Yes/No) for each customer, the data contains information on provided services in variables Est_Income , Car_Owner , Usage , Age group , RatePlan , LongDistance ,International Calls ,Local Calls, Dropped Calls. It also contains customer account information in variables Paymethod, LocalBilltype, LongDistanceBilltype. Finally, customer demographic information is contained in variables gender, Status, Dependents. Purpose The purpose of this study is, with the help of Watson Analytics examine why customers are not used the connection of Bits Telecom Company, which factors are influence the churn. Also see the cross selling and up-selling, also focus on profitability and investment and find out the way for better results.
- 3. Task 1 – CHURN ANALYSIS 1.1– With the help of the Watson Analytics first do Exploration of data. When Explore the data some starting points are Suggest by Watson. Purpose-The purpose of this section to find out what factors influence churn and why
- 4. What the breakdown of the number of contract by Churn? Explore the data of contract by churn,breakdown of…in the search box and at the window the most relevent options are shows.now choose one option and click on it. According to the company’s contracts the churn equally distributes. -We can add new columns by the option “add new column”
- 5. Add a column Age group (6 type of age group) By the visualization option can change the visualization. The area chart shows that all the age groups are interested in company’s contracts, but according to churn the contracts are equally popular in all the age groups in cancelled customers and in the current customers the popularity of contracts in two age groups is less, but all over users are equal to canceled. Change the visualization - choose the bar chart.
- 6. - This chart shows clearly that the two year contract is used by only 4 age groups and month to month and one month contract are again at equal level. Focus on Two year contract because the current customers of this contract are not greater then cancelled customers. Change the visualization(Line Chart) -change the visualization option. -Pin the chart. - This chart shows the variation of the rate plan popularity between current and cancelled customers.(red line)current customers, only month to month contract raise other two are highly decrease but overall the current customers whom used the rate plans are very less then to cancelled. First focus on Two year rate plans. -Three types of customers. The status are- Married Single Divorced
- 7. Which category is interested in two year rate plans - - Only 13 customers are using the Two year plan others already canceled it. The Est_Income by the two year plan. Filter – two filters apply here, that means once you choose any field as a filter then furthers all the Exploration proceed according to that filter. Here I select Contract (Two year) because we analyze the two year contract popularity and other is rate plan (all). - Choose the filter option. - Add a filter. - Conditions.
- 8. - The Est_income is only 319.7k by the current customers and by the canceled customers 24.67m. the conclusion is because of churning the company has vey big loss. There are 4 type of rate plans in company - Which rate plan is used broadly in Two year contract? - In the line chart, the current customers only use RT1, RT3, RT4 and canceled customers used all 4 plans. The current customers are very less.
- 9. 1.2 Prediction – Which fields are drives the churn. All the factors, which affect the churn The Top Predictors of Churn spiral chart shows the Combination model and various Two Field models that are more predictive of customer churn than One Field models .The predictive strength of one field is 73% shows by the BAR chart and the driving fields are contract and dependents. The combination of the fields, the predictive strength is 74% and driving factors is (contract and age group), (age group and dropped) shows by the BAR chart…
- 10. There are three types of contract in company Month to Month, One year, Two year In area chart, according to age group churn position is define. Green and Blue color area define (yes/no). As the record the month to month contract has used more then one year and two year contract. The age group 36-45 used this contract broadly and greater number of current customer who is used month to month pack. Churn ratio going down by one year and two year contract and very less number of current customers interested in two year contract. Now focus on Two year contract, which rate plan is used mostly in two year contract? We should modify the two year contract.
- 11. If we see the rate plan history then by line chart it is clear that the scope of success of rate plan 4 and but at the current time only three rate plan is used in two year contract so the condition is very poor at that time. First talk about the rate plan 4 applies some discount and offers for the users. You can add some policy to attract the customers for use that service broadly. For example check which status categories use the Two year plan and their dependents. The divorce status not uses the Two year plan and haven’t dependents and The Married and single customers used it and they have dependents.
- 12. Now for example can offer them like “with one connection another connection free” etc. Other hand for divorced can offer like discount policy. 1.3 Assemble -Create the dashboard for the clear view of data and create a story. To create a dashboard, three types of templates here Single Page, Tabbed, Infographics. (Tabbed->Freeform->create)
- 13. What do you want to assemble, write in search box then it is create in dashboard. There are 4 options for a single view box- Delete, Duplicate-to create a duplicate tab, Edit Title, Visualization.
- 14. Select column chart Dashboard properties- Themes- apply a theme. General Style- Change the style. Pinning Service – Pin (save) the tab.
- 15. Slide Show Select BETA slide show option. Select a Template. Create. Open pin Charts. Drag the chart at tab.it will be create at dashboard.
- 16. Adjust the chart in the dashboard. Allow full screen. Conclusion is, in the two year contract, very few customers interested. It is affecting the churn. So focus on two year contract and in this contract, the Rate Plans, to increase the customers.
- 17. Task – 2 Cross-Selling and Up-Selling 2.1 EXPLORE the data, What is the breakdown of Usage by Long distance bill type and local bill type? (In the long distance bill type 2 categories are Standard and International discount and in Local bill type 2 categories are Budget (postpaid) and free local (prepaid)). According the data the international discount bill type Customers are use this services not as much of. The other category used broadly. The purpose of this section is to create new ideas and How to implement those ideas, for the better Income and selling. In this section we focus on Bill types and usage of the customers and analyze that how to get more income in business.
- 18. Now focus on International discount. Add a filter (LongDistanceBilltype -> Select International Discount) The average of total usage, 16.13% utilize by free local bill type service and 28% utilize by Budget. Means, Budget category use the service broadly. Now see the difference according the calls type.
- 19. Add another filter “Usage”. Three types of calls system are present in company. Local Calls Long distance Calls International Calls Here we analyze that which type of calls are more utilize and which payment method they use. Three facilities of payment CC (Credit card) CH (Check) Auto (Cash) Local Calls Budget (Postpaid) bill type Category utilize Local calls higher then Free local (Prepaid).
- 20. And the pay methods of this category. Add 3 filter local bill type, select free local (prepaid). They are use Auto and CC mostly and CH is not much popular. So, for this category can offer some policy to CC company to increase profit and CH category like banks here we can offer the banks others beneficial plans for example if bank offer their customers some plans including our company policy to promote telecom company and also with CH facilities then we provide the connections free of cost or with discount. Long Distance Calls
- 21. Almost use by both category. Now focus on free local (prepaid). The pay methods of this category. Add 3 Filter Local Bill Type, select Free local (prepaid). Auto (Cash) method is popular, should add some plan in top ups. For example- When Customers buy the pack then offer them same type of pack with extra balance or 3G balance but price is high. CH (check), collaborate with banks and apply some other policy to promote this service. International Calls
- 22. Postpaid Connection Customers utilize international calls broadly and prepaid customers are not use to much. Now focus on free local type (prepaid). The pay methods of Free local Bill type customers. Add 3 filter Local Bill Type, select Free local (prepaid). Only CC customers use broadly the service. Auto customers very less, offer some discount, if they choose cash payment option. CH (check), collaborate with banks and apply some other policy to promote this service. Conclusion is, Prepaid Connection is not using the services broadly. In the Payment methods we should focus and then add some policy to promote the company.
- 23. 2.2 PREDICTION The main focus of the business owner at income. Select the target. Top Predictors of Estimated income are Local Calls and 8 others. Shows in decision tree.
- 24. Top Decision rules, predicting Estimated income to be high. The Local calls highly influent the income. Also long distance calls and local bill type, as seen in exploration part should focus on these factors.
- 25. 2.3 ASSEMBLE Create a dashboard and story. Create full story of analysis for clear view.
- 26. Task 3 - Customer Segmentation & Customer Lifetime Value Analytics 3.1 EXPLORE the data, If divide the customers age group according to income. The estimation is that highest income by 16-25 age group customers. Customers are base of the any business and for the better result the long term business relationship with customer is important. In this section we focus on customer’s priority according their involvement in our company also plan some offers for the users whom used our service from long time according their needs and status.
- 27. Now observe the dependents of the particular age group. As shows in tree map chart that age group which has no dependents, the income is high and other which have dependents, the income is less. So, focus on that group which have dependents here if offer that groups some kind of plan like “connection free” or some application like, “at one connection discount for other connection”. Conclusion is, the customers with dependents, here better scope of selling company’s plans.
- 28. Other hand, the income is high by customers who haven’t car. What is the breakdown of estimated income by car owner? Conclusion is, these customers are very important for company here we should offer them some special offers, it is beneficial for company. The car owners (think that they can afford some expensive offers so, can offer them some expensive plans with long validity).
- 29. 3.2 PREDICTION – Top predictors of car owner – the top predictor is estimated income and 6 other input. Decision tree shows all inputs.
- 30. Highest decision rules for that that haven’t car with statistical details. Highest Decision rules for who have car.
- 31. Top predictor is estimated income for car owner as discussed in exploration part we should focus on both type of customers. 3.3 ASSEMBLE – create Slide Show. Conclusion is, Income is high by those customers whom haven’t car. They are important customers for company so we should offer them some good plans.