Merge churn propensity into pCLV; retire the old predictive-UI churn page - #871
Draft
jmruggiero wants to merge 3 commits into
Draft
Merge churn propensity into pCLV; retire the old predictive-UI churn page#871jmruggiero wants to merge 3 commits into
jmruggiero wants to merge 3 commits into
Conversation
PR #862 rewrote the EPM model and overview pages but left several other files describing the obsolete manual setup flow, and EPM pages were surfacing shared snippets written for product affinity / purchase models. - table_event_propensity.rst: replace the manual "add the table" SQL flow (which joined Predicted_Affinity_* passthrough tables, copied from the Product Affinity table page) with an explanation that the model auto-creates Predicted_Propensity_{EventName}. Remove a stray Unified Product Catalog note that never applied to EPM. - data_tables.rst: correct the Event Propensity Score field -- ranking is derived directly from the score, so drop the "does not directly correlate to ranking" claim while keeping the steer to target on Ranking / audience size (the score is uncalibrated). Reword the Event Propensity audience-size columns from "future purchasers" to "customers who are likely to perform the target event". - terms.rst: add an event-worded term-recommended-audience-size-events variant (the shared term is product-affinity-specific) and point the three EPM pages at it; de-duplicate the malformed event-propensity-table term block. - Correct the output/how-tos wording: users only need to activate an EPM model (not activate, schedule, and run). - Minor EPM grammar fixes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Churn propensity is an output of the predicted CLV (pCLV) model -- the UI model-type picker offers no separate "churn propensity" model, and the pCLV card itself is described as predicting "value tier and churn risk" (model_type.cljs). The churn build section still documented the old single-page "Predictive enablement" flow (model start date, table selection, prediction horizon, additional features, "Start validation"), which no longer exists. Replace those steps with a short section that explains churn attributes are produced by the pCLV model and points to "Build a predicted CLV model" (which is already documented on the current UI). The pCLV and product affinity build docs were already updated to the new flow; churn was the last predictive-model page on the old UI. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Churn propensity is not a separate model -- it is an output of the predicted CLV (pCLV) model (the UI has no churn model type; the pCLV model itself predicts "value tier and churn risk"). Maintaining separate churn "model" pages caused duplication and drift. - Fold the churn-specific content (retention framing, predicted customer lifecycle status tiers) into a new "Which customers are likely to churn?" use case on the pCLV pages (operator/reference/user). The reusable block lives in the reference pCLV page and is included by the operator and user pages. - Delete the standalone churn pages: model_churn_propensity.rst (operator, reference) and model_predicted_churn.rst (user). - Remove the "Churn propensity" model cards and toctree entries from the model grids and the user index; drop the churn overview section and build how-tos from the operator models overview, folding a churn mention into the pCLV overview. - Reword the churn propensity glossary term to frame it as a pCLV output. NOTE: URL redirects for the removed pages (operator/reference model_churn_propensity.html, user model_predicted_churn.html -> the pCLV pages) must be configured at the CDN/server level -- this repo has no in-tree redirect mechanism. make operator/reference/user all build clean with -W. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
jmruggiero
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Stacked on #870 (EPM). This started as "clean up the churn page's obsolete setup flow" and, per follow-up discussion, became a full information-architecture fix: churn propensity is not a separate model — it is an output of the predicted CLV (pCLV) model — so it no longer has its own "model" pages.
Evidence it's not a separate model: the UI model-type picker offers only pCLV, product affinity, and the EPM presets, and the pCLV card itself is described as predicting "value tier and churn risk" (
model/create/model_type.cljs). pCLV and product affinity docs were already on the current UI; churn was the last page describing the old "Predictive enablement" / "Start validation" flow, and it duplicated pCLV.Changes
Content merge
Deletions
amperity_operator/source/model_churn_propensity.rstamperity_reference/source/model_churn_propensity.rstamperity_user/source/model_predicted_churn.rstNavigation / cross-refs
:doc:/:ref:links.Net: 12 files, +59 / −550.
This repo has no in-tree redirect mechanism (no rediraffe/reredirects, no meta-refresh stub convention —
_publicholds only assets). URL redirects for the removed pages must be set up at the CDN/server layer:/operator/model_churn_propensity.html→/operator/model_predicted_clv.html#...churn/reference/model_churn_propensity.html→/reference/model_predicted_clv.html#...churn/user/model_predicted_churn.html→/user/model_predicted_clv.html#...churnI can add meta-refresh HTML stubs under each guide's
_public/if you'd prefer an in-repo stopgap — just say the word.Validation
make operator,make reference,make userall build clean with-W. Verified the churn use case renders on all three pCLV pages (lifecycle tiers included) and the deleted pages are gone.🤖 Generated with Claude Code