Cltv Prediction

Copy

• Copy for LLM

  Copy page as Markdown for LLMs
• View as Markdown

  Open this page as Markdown
• Open in ChatGPT

  Get insights from ChatGPT
• Open in Claude

  Get insights from Claude
• Connect to Cursor

  Install MCP server on Cursor
• Connect to VS Code

  Install MCP server on VS Code

The CLTV Prediction solution notebook forecasts customer lifetime value (CLTV) using an open source Python library Lifetimes for implementing the Buy Till You Die (BTYD) model.

BTYD is implemented using:

• The BG/NBD (Beta-Geometric/Negative Binomial Distribution) model fits on transactional history curves describing the distribution of purchase frequencies and engagement drop-off following a prior purchase.
• The Gamma-Gamma model is used to predict average spend per transaction.

Expected Input

This notebook takes the same input format to RFM analysis notebook, a transaction table specified by input_table option, having user_column , tstamp_column , and amount_column columns, as follows:

Treasure Data accepts various ISO-8601 datetime format supported by dateutil as well as unix timestamp for the tstamp column.

Expected Outcome

In output_table:

• Note that frequency, recency, and tenure are measured in days.
• The tenure represents the number of days since the customer initially issued a transaction.
• CLTVs (monetary values) for the next 1, 3, 6, 12, 24 months are exported.
• Percentile rank takes value between 0 and 100.
• The automl_alive_prob represents probability of customers alive now.

When audience_name is specified as an input parameter, the following values will be created as new attributes of the specified parent segment where Xm represents months later.

• automl_cltv_Xm
• automl_cltv_Xm_pctile
• automl_cltv_segment
• automl_alive_prob

The automl_cltv_segment splits automl_cltv_12m into five groups (very low/low/medium/high/very high) using quantiles, and then generates five CDP Segments.

Example Graph Outputs:

Workflow Example

Find a sample workflow in Treasure Boxes.

+run_cltv:
  ipynb>:
    notebook: CLTV
    input_table: ml_datasets.online_retail_txn
    output_table: ml_results.online_retail_cltv_result
    user_column: customerid
    tstamp_column: invoicedate
    amount_column: purchaseamount

Parameters