Customer activation Predictive model
- 2. Customer Activation Focus on Equity Objectives 1. To predict activity levels of each customer in the near future( (Current Model: 90 days) 2. To profile customer activity over time (i.e., activity states with durations) 3. To determine the recommendations to activate customers Problem Dimensions • People – Who are the people likely to be inactive in the next month? • Activity State – What are the different states in customer life cycle? – What is the customer behaviour in a particular state? • State Duration – How long would the customer will be in particular state? – What will be the transition time for a particular customer? • Recommendation – What strategy will be effective to prohibit inactivity of a particular customer? – What strategy can bring customer back from inactive state to active state?
- 3. Analysis Process Distributions: Inactive period behaviour, life cycle of customer Comparative views: First time inactive vs. current inactive, inactive vs. active customer life cycle ETL Merge Filter Visualize Storage: ACMIIL (Trades) Data Formats: Dates, categories, numeric value ranges, etc. File Formats: Comma, Tilde, or Tab delimited Customer types: Individual vs. Institutions, etc. Transaction types: Buying/Selling, First time inactive, current inactive Identifiers: Client Code, CommonClient Code Timeline: Daily, Monthly Aggregates: Counts, Sums of EQ buy, Sums of EQ sell
- 4. Activity Modelling - Outline Trades Data • Summary • Discovery Model • Setup • Application • Code • Results • Setup - Next Steps • Application – Next Steps Future Work
- 5. Data: Summary Statistical measures (e.g., mean) errors – Units field has negative values – Too large or small values Text data: Mis-matches Numerical data: Unreal ranges Numerical data: Spurious values DQ Issues Sizing for technology • ~7M EQ and ~1M DER trades per year • ~100k trading customers currently on platform, and 1/3rd transacted in the last 6 months Analysis caution • Data distributions highly skewed, e.g., few high amount Txs by one or two individuals
- 6. Data: Discovery Inactivity count All clients inactive at least once for greater than 91 days Inactivity Count Insights All Clients have been inactive at least once Frequency Data Particulars Data Duration 2012 Apr - 2015 Sep Each Row Client-month Client Category Individual and HUF # of Rows 366421 # of Columns 31 # of Unique Clients 49444
- 7. Data: Discovery Average Inactive Duration Average inactivity duration (days) Frequency 300 days Insights Histogram of Average inactivity duration gives maximum frequency at 300 days Data Particulars Data Duration 2012 Apr - 2015 Sep Each Row Client-month Client Category Individual and HUF # of Rows 366421 # of Columns 31 # of Unique Clients 49444
- 8. Data: Discovery First Time Inactive vs. Currently Inactive First time inactive Currently inactive Vintage (yrs) Vintage (yrs) Frequency Frequency 5 yrs 7 yrs Insights Current inactive customers are a mix of first time inactive and other periods making it harder to study current inactivity alone => It brings about the need to study each activity level or state separately Data Particulars Data Duration 2012 Apr - 2015 Sep Each Row Client-month Client Category Individual and HUF # of Rows 366421 # of Columns 31 # of Unique Clients 49444
- 9. SumAmt.(sold)SumAmt.(Bought) Data: Discovery Random customer 1: currently inactive(Tx Amount) Trend Curve Trend Curve
- 10. Data: Discovery Random customer 1: currently Inactive (Tx Count) TxCount(sold)TxCount(bought) Trend Curve Trend Curve
- 11. Data: Discovery Random customer 2: currently active(Tx Amount)SumAmt.(sold)SumAmt.(Bought) Trend Curve Trend Curve
- 12. Data: Discovery Random customer 2: currently active (Tx Count) TxCount(sold)TxCount(bought) Trend Curve Trend Curve
- 13. Data: Discovery Insights • All clients have been inactive (> 91 days inactivity) at least once • The most-likely inactivity duration is ~300 days, i.e., if customer becomes inactive => there is a high chance of a long inactivity period • Customer behaviour is different before various inactive states • Each inactive state (i.e., first time or second time, etc.) need to be modelled separately • There are different trend curves in a customer’s life cycle that each of customers follow • The trend curves may be grouped together into a finite set of representative trend curves • All the above may be modelled using a State-space approach • A simple binary approximation is the Logistic regression model
- 14. Test Data Three Year Trade Data 60% Used for Training Model 20% Used for Validating Model 20% Used for Testing Model Total Available Data Training Data Validation Data Time Acc Opening Date 1 1 First Time inactive Inactive 1 Active Period Inactive Period Inactivity: Defined as 0 transactions in consecutive 91 days Hypothesis: Customer’s state can be predicted using transactions data Logistic Regression Model To find predictive variables To predict next state of the customer 0 0 0 0 Data Set Creation Model: Setup
- 16. Model: Application 0 0 1 0 0 0 1 0 0 0 0 1 0 1 Actual States Predicted States Inactive State miss Active State miss Actual Predicted Positive Positive Negative Negative a b c d a - True Positive b - False Negative c - False Positive d - True Negative 𝐻 𝑎 = 𝑑 𝑁0 𝑀 𝑎 = 𝑏 𝑁0 𝐻𝑖 = 𝑎 𝑁1 𝑀𝑖 = 𝑐 𝑁1 𝐻 𝑎 - Active state hit rate 𝑀 𝑎- Active state miss rate 𝐻𝑖 - Inactive state hit rate 𝑀𝑖 - Inactive state miss rate
- 17. Model: Results 𝑀 𝑎= 0.01% 0 5000 10000 15000 20000 25000 30000 35000 Correct Predicted Active State Wrong Predicted Active State 0 5000 10000 15000 20000 Correct Predicted Inctive State Wrong Prdicted Inctive State 𝐻𝑖 = 84.5% Threshold = 0.25 0 5000 10000 15000 20000 25000 30000 35000 Correct Predicted Active State Wrong Predicted Active State 𝑀 𝑎= 60.8% 0 5000 10000 15000 20000 25000 Correct Predicted Inctive State Wrong Prdicted Inctive State 𝐻𝑖 = 93.1% Threshold = 0.35 0 10000 20000 30000 40000 50000 60000 Correct Predicted Active State Wrong Predicted Active State 0 5000 10000 15000 20000 25000 Correct Predicted Inctive State Wrong Prdicted Inctive State 𝑀 𝑎= 40.3% 𝐻𝑖 = 0.0009% Threshold = 0.50 𝑀 𝑎- Active state miss rate 𝐻𝑖 - Inactive state hit rate a a a c c c d d db b b
- 18. Model: Application (next steps) Multi-period Hypothesis: - Error rates can be decreased by taking into account multiple periods for predictions 0 0 1 0 0 0 1 0 0 0 0 1 0 1 Actual States Predicted States Model predicts 1 Check customer’s transaction in next 30 days If Tx = 0 Model output is 0 Model output is 1 TrueFalse 1 Active Period Inactive Period 0
- 20. active inactive closed On- boarded Technical Model: State-space Model • In the applied model we have taken only two states 0 for active and 1 for inactive • Between these active and inactive state a customer can transit into many different states as shown in the state space model above • By applying state space model the complete life cycle of a customer i. Previous state ii. Next state iii. Time he will be in a particular state iv. Behaviour of customer in a particular state v. Behaviour of customer just before transition, vi. Behaviour of customer before going off-board, etc., will be profiled
- 22. Back-up Slides
- 24. Model: Setup (next steps) Customer Sampling For the current model, Training, validation and Testing dataset has been created by sampling on the basis of rows, where each row is a particular customer and aggregated transaction amounts on monthly basis. We can create Training, validation and Testing dataset by sampling as per customer basis.