Clean up cohort normalization math

.pre-commit-config.yaml

@@ -1,24 +1,26 @@
 repos:
   - repo: https://github.com/dbt-checkpoint/dbt-checkpoint
-    rev: v1.0.0
+    rev: v1.1.0
     hooks:
       - id: dbt-docs-generate
       - id: check-script-semicolon
         files: ^models
       - id: check-model-has-properties-file
       - id: check-script-has-no-table-name
         files: ^models
-      - id: unify-column-description
+      - id: generate-model-properties-file
+        args: ["--properties-file", "/models/{schema}/{name}.yml", "--"]
+      # - id: unify-column-description
       # - id: check-column-name-contract
       #  args: [--pattern, "(is|has|do)_.*", --dtype, boolean]
       - id: check-model-has-all-columns
-        name: Check columns - mart
-        files: ^models/mesa/mart
+        name: Check columns - marts
+        files: ^models/mesa/marts
       - id: check-model-has-tests
-        files: ^models/mesa/mart
+        files: ^models/mesa/marts
         args: ["--test-cnt", "1", "--"]
-      - id: check-model-columns-have-desc
-        files: ^models/mesa/mart
+      # - id: check-model-columns-have-desc
+      #  files: ^models/mesa/marts
 
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.4.0
@@ -28,6 +30,6 @@ repos:
       - id: trailing-whitespace
         exclude: ^.bumpversion.cfg
   - repo: https://github.com/pre-commit/mirrors-prettier
-    rev: "v3.0.0-alpha.4"
+    rev: "v3.0.0-alpha.6"
     hooks:
       - id: prettier

README.md

@@ -153,7 +153,7 @@ The application's models are structured into 3 types & folders.
 
 - staging: The initial import of raw source data with filtering and slight formatting.
 - intermdiate: intermediary formatting and supporting models that are not business user-facing.
-- mart: The finalized production-ready business insight models.
+- marts: The finalized production-ready business insight models.
 
 ## Oddities & Gotchas
 

dbt_project.yml

@@ -84,7 +84,7 @@ models:
           # +on_schema_change: "sync_all_columns"
           # +unique_key: shop_subdomain
     mesa:
-      mart:
+      marts:
         step_runs:
           +on_schema_change: "sync_all_columns"
           +cluster_by: ["shop_subdomain"]
@@ -145,6 +145,12 @@ models:
       staging:
         +schema: support
 
+  dbt_project_evaluator:
+    marts:
+      structure:
+        fct_model_naming_conventions:
+          +enabled: false
+
 tests:
   shoppad:
     mesa:

models/generic/staging/stg_constellation_users.sql

@@ -2,6 +2,9 @@ WITH constellation_users AS (
     SELECT
         COALESCE(uuid, id) AS shop_subdomain,
         createdat AS first_in_constellation_at_utc,
+        {{ pacific_timestamp('first_in_constellation_at_utc') }} AS first_in_constellation_at_pt,
+        first_in_constellation_at_pt::DATE AS first_in_constellation_on_pt,
+        date_trunc('week', first_in_constellation_on_pt) AS constellation_cohort_week,
         COALESCE(updatedat, createdat, uuid_ts) AS updated_at,
         analytics_gmv,
         analytics_orders,

models/generic/staging/stg_shop_infos.sql

@@ -18,6 +18,7 @@ WITH flattened_table AS (
 
     FROM {{ source_table }}
     WHERE uuid IS NOT NULL
+        AND DATA != '[]'
     QUALIFY ROW_NUMBER() OVER (PARTITION BY shop_subdomain ORDER BY created_at DESC) = 1
 ),
 

models/mesa/mart/_shops.yml → models/mesa/marts/_shops.yml

@@ -438,3 +438,15 @@ models:
         description: The timestamp of the Shop's last email conversion.
       - name: first_in_constellation_at_utc
         description: The timestamp of the Shop's first install in the constellation.
+      - name: first_in_constellation_on_pt
+        description: The date of the Shop's first install in the constellation apart from MESA.
+        meta:
+          metabase.display_name: First In Constellation On (PT)
+      - name: first_in_constellation_at_pt
+        description: The timestamp of the Shop's first install in the constellation apart from MESA.
+        meta:
+          metabase.display_name: First In Constellation At (PT)
+      - name: constellation_cohort_week
+        description: The week of the Shop's first install in the constellation apart from MESA.
+        meta:
+          metabase.display_name: Constellation Cohort Week

models/mesa/mart/_workflows.yml → models/mesa/marts/_workflows.yml

@@ -133,3 +133,5 @@ models:
         description: Whether the workflow has been deleted by the Shop.
         tests:
           - not_null
+      - name: setup
+        description: The stage of a template-wizard workflow setup process.

models/mesa/mart/rollups/_mesa_shop_days.yml → models/mesa/marts/rollups/_mesa_shop_days.yml

@@ -118,3 +118,9 @@ models:
         description: Whether the shop completed the wizard on this day.
       - name: workflows_created_without_wizard_count
         description: The number of workflows created without using the wizard on this day.
+      - name: is_in_trial
+        description: Whether the shop is in a trial on this day.
+      - name: is_shopify_zombie_plan
+        description: Whether the shop is on a zombie plan on this day.
+      - name: mesa_plan_identifier
+        description: The Mesa plan identifier for the shop on this day.

models/mesa/mart/rollups/_mesa_shop_monthly_cohort_performance.yml → models/mesa/marts/rollups/_mesa_shop_monthly_cohort_performance.yml

@@ -373,3 +373,23 @@ models:
         description:
           The percentage of users who first installed in the given month and
           have enabled a workflow within the first six months.
+      - name: shopify_unlimited_plan_count
+        description: The number of users who have a Shopify Unlimited plan.
+      - name: shopify_basic_plan_count
+        description: The number of users who have a Shopify Basic plan.
+      - name: shopify_trial_plan_count
+        description: The number of users who have a Shopify Trial plan.
+      - name: shopify_unlimited_plan_pct
+        description: The percentage of users who have a Shopify Unlimited plan.
+      - name: shopify_professional_plan_pct
+        description: The percentage of users who have a Shopify Professional plan.
+      - name: shopify_shopify_plus_plan_pct
+        description: The percentage of users who have a Shopify Plus plan.
+      - name: shopify_professional_plan_count
+        description: The number of users who have a Shopify Professional plan.
+      - name: shopify_trial_plan_pct
+        description: The percentage of users who have a Shopify Trial plan.
+      - name: shopify_basic_plan_pct
+        description: The percentage of users who have a Shopify Basic plan.
+      - name: shopify_shopify_plus_plan_count
+        description: The number of users who have a Shopify Plus plan.

models/mesa/mart/rollups/_mesa_shop_weekly_cohort_performance.yml → models/mesa/marts/rollups/_mesa_shop_weekly_cohort_performance.yml

@@ -373,13 +373,13 @@ models:
         description:
           The percentage of users who first installed in the given week and
           have enabled a workflow within the first six months.
-      - name: non_mesa_shop_count
+      - name: last_period_avg_constellation_cohort_size
         description: The number of new installed Shops but are not in the Mesa
           segment.
       - name: shopify_shopify_plus_plan_count
         description: The number of users who have a Shopify Plus plan.
-      - name: non_mesa_mql_count
-        description: The number of MQLs who are not in the Mesa segment.
+      - name: avg_constellation_mql_count
+        description: The number of MQLs who are not in the Mesa segment averaged over trailing 4 weeks.
       - name: shopify_trial_plan_count
         description: The number of users who have a trial plan.
       - name: shopify_trial_plan_pct
@@ -392,37 +392,39 @@ models:
         description: The percentage of users who have a Basic plan.
       - name: shopify_basic_plan_count
         description: The number of Mesa users who have a Basic plan.
-      - name: one_week_non_mesa_mql_growth_rate
+      - name: avg_constellation_mql_count_growth
         description: The percentage of non-MESA MQL shops in the constellation for the given week compared to the week prior.
-      - name: prior_one_week_non_mesa_customer_count
-        description: The number of non-MESA new customer app installs in the prior week.
+      - name: last_period_avg_constellation_cohort_size
+        description: The rolling average number of non-MESA new customer app installs in the prior week.
       - name: shopify_unlimited_plan_pct
         description: The percentage of users who have an Unlimited plan.
-      - name: one_week_mesa_growth_rate
+      - name: mesa_growth_rate
         description: The percentage of new customer MESA installs for the given week compared to the week prior.
-      - name: one_week_mql_growth_rate
+      - name: mql_growth_rate
         description: The percentage of new customer MQL MESA installs for the given week compared to the week prior.
       - name: shopify_unlimited_plan_count
         description: The number of users who have an Unlimited plan
       - name: prior_one_week_mql_count
         description: The number of MQLs in the prior week.
-      - name: prior_one_week_non_mesa_mql_count
-        description: The number of non-MESA MQL shops in the constellation in the prior week.
+      - name: avg_constellation_cohort_size
+        description: The number of non-MESA MQL shops in the constellation averaged over the last 4 weeks.
       - name: shopify_shopify_plus_plan_pct
         description: The percentage of users who have a Shopify Plus plan.
       - name: mql_count
         description: The number of MQLs that installed MESA in the given week.
-      - name: one_week_non_mesa_growth_rate
-        description: The percentage of non-MESA new customer app installs for the given week compared to the week prior.
+      - name: avg_constellation_cohort_size_growth
+        description: The percentage of non-MESA new customer app installs averaged over the last 4 weeks.
       - name: prior_one_week_new_customer_count
         description: The number of new customer MESA app installs in the prior week.
-      - name: one_week_normalized_customer_count
+      - name: normalized_customer_count
         description: The number of new customer MESA installs for the given week normalized by the number of growth rate of the constellation.
-      - name: normalized_growth_rate
+      - name: normalized_customer_growth
         description: The grwoth rate of MESA normalized by the growth rate of the constellation.
-      - name: double_normalized_growth_rate
-        description: The MESA normalized insall count multipled by the normalized growth rate of the constellation.
-      - name: prior_one_week_normalized_mql_count
-        description: The number of MQLs in the prior week normalized by the number of growth rate of the constellation.
-      - name: one_week_normalized_mql_count
+      - name: normalized_mql_count
         description: The number of MQLs in the given week normalized by the number of growth rate of the constellation.
+      - name: last_period_avg_constellation_mql_count
+        description: Last week's number of MQLs who are not in the Mesa segment averaged over trailing 4 weeks.
+      - name: avg_constellation_mql_pct
+        description: The percentage of MQLs who are not in the Mesa segment averaged over trailing 4 weeks.
+      - name: normalized_mql_growth
+        description: The growth rate of MQLs normalized by the growth rate of the constellation.

models/mesa/mart/rollups/_monthly_cohort_performance.yml → models/mesa/marts/rollups/_monthly_cohort_performance.yml

@@ -4,52 +4,52 @@ models:
   - name: monthly_cohort_performance
     description: "This table breaks paying shops into cohorts based on when they starrted to pay and into monthly revenue performance"
     tests:
-        - dbt_expectations.expect_compound_columns_to_be_unique:
-            column_list:
-                - first_month
-                - active_month
+      - dbt_expectations.expect_compound_columns_to_be_unique:
+          column_list:
+            - first_month
+            - active_month
     columns:
       - name: first_month
         description: "The first month the Shop paid us."
         tests:
-            - not_null
+          - not_null
 
       - name: active_month
         description: "The month of the performance."
         tests:
-            - not_null
+          - not_null
 
       - name: months_since_first
         description: "The number of the period after the first paying month."
         tests:
-            - not_null
+          - not_null
 
       - name: shops
         description: "The count of Shops that paid this month."
         tests:
-            - not_null
+          - not_null
 
       - name: cohort_num_users
         description: "The total count of the cohort's Shops at the beginning"
         tests:
-            - not_null
+          - not_null
 
       - name: retained_pctg
         description: "The percent of Shops that paid this month compared to the total cohort."
         tests:
-            - not_null
+          - not_null
 
       - name: inc_amt
         description: "The revenue of this cohort in this month."
         tests:
-            - not_null
+          - not_null
 
       - name: cum_amt
         description: "The sum total of all cohort revenue up through this period."
         tests:
-            - not_null
+          - not_null
 
       - name: cum_amt_per_user
         description: "The average revenue per Shop in this cohort up to this period."
         tests:
-            - not_null
+          - not_null

models/mesa/mart/rollups/mesa_shop_weekly_cohort_performance.sql → models/mesa/marts/rollups/mesa_shop_weekly_cohort_performance.sql

@@ -147,17 +147,45 @@ mql_counts AS (
     GROUP BY 1
 ),
 
-non_mesa_installs AS (
+weekly_constellation_installs AS (
     SELECT
-        date_trunc('week', updated_at) AS cohort_week,
-        COUNT(*) AS non_mesa_shop_count,
-        COUNT_IF(is_mql) AS non_mesa_mql_count
-    FROM {{ ref('int_shop_infos') }}
+        constellation_cohort_week,
+        COUNT(stg_constellation_users.*) AS constellation_cohort_size,
+        COUNT_IF(stg_constellation_users.is_mql) AS constellation_mql_count
+    FROM {{ ref('stg_constellation_users') }}
     LEFT OUTER JOIN shops USING (shop_subdomain)
     WHERE shop_subdomain NOT IN (SELECT * FROM {{ ref('staff_subdomains') }})
     GROUP BY 1
 ),
 
+constellation_comparison_groups AS (
+    SELECT
+        constellation_cohort_week AS cohort_week,
+        AVG(constellation_cohort_size)
+            OVER (ORDER BY constellation_cohort_week ROWS BETWEEN 3 PRECEDING AND CURRENT ROW)
+            AS avg_constellation_cohort_size,
+        AVG(constellation_mql_count)
+            OVER (ORDER BY constellation_cohort_week ROWS BETWEEN 3 PRECEDING AND CURRENT ROW)
+            AS avg_constellation_mql_count,
+        avg_constellation_mql_count / avg_constellation_cohort_size
+            AS avg_constellation_mql_pct
+    FROM weekly_constellation_installs
+),
+
+decorated_constellation_comparison_groups AS (
+    SELECT
+        *,
+        LAG(avg_constellation_cohort_size)
+            OVER (ORDER BY cohort_week) AS last_period_avg_constellation_cohort_size,
+        LAG(avg_constellation_mql_count)
+            OVER (ORDER BY cohort_week) AS last_period_avg_constellation_mql_count,
+        avg_constellation_cohort_size / last_period_avg_constellation_cohort_size - 1
+            AS avg_constellation_cohort_size_growth,
+        avg_constellation_mql_count / last_period_avg_constellation_mql_count - 1
+            AS avg_constellation_mql_count_growth
+    FROM constellation_comparison_groups
+),
+
 shopify_plan_counts AS (
     SELECT
         cohort_week
@@ -178,37 +206,31 @@ final AS (
         total_ltv_revenue / NULLIF(has_enabled_a_workflow_count, 0) AS lifetime_value_enabled_workflow,
         total_ltv_revenue / NULLIF(is_activated_count, 0) AS lifetime_value_activated,
 
-        {# Rest of Constellation Growth #}
-        LAG(non_mesa_shop_count) OVER (ORDER BY cohort_week) AS prior_one_week_non_mesa_customer_count,
-        1.0 * cohort_size / NULLIF(prior_one_week_non_mesa_customer_count, 0) -1 AS one_week_non_mesa_growth_rate,
-
-        {# Rest of Constellation MQL Growth #}
-        LAG(non_mesa_mql_count) OVER (ORDER BY cohort_week) AS prior_one_week_non_mesa_mql_count,
-        1.0 * non_mesa_mql_count / NULLIF(prior_one_week_non_mesa_mql_count, 0) - 1 AS one_week_non_mesa_mql_growth_rate,
 
         {# MESA Install Growth #}
         LAG(cohort_size) OVER (ORDER BY cohort_week) AS prior_one_week_new_customer_count,
-        1.0 * cohort_size / NULLIF(prior_one_week_new_customer_count, 0) - 1 AS one_week_mesa_growth_rate,
-        1.0 * prior_one_week_non_mesa_customer_count / (1 + NULLIF(one_week_non_mesa_growth_rate, 0)) AS one_week_normalized_customer_count,
+        cohort_size / NULLIF(prior_one_week_new_customer_count, 0) - 1 AS mesa_growth_rate,
+        cohort_size * NULLIF(1 - avg_constellation_cohort_size_growth, 0) AS normalized_customer_count,
+        mesa_growth_rate * NULLIF(1 - avg_constellation_cohort_size_growth, 0) AS normalized_customer_growth,
 
         {# MESA MQL Growth #}
         LAG(mql_count) OVER (ORDER BY cohort_week) AS prior_one_week_mql_count,
-        1.0 * mql_count / NULLIF(prior_one_week_mql_count, 0) - 1 AS one_week_mql_growth_rate,
-        1.0 * mql_count / NULLIF(1 + one_week_non_mesa_mql_growth_rate, 0) AS one_week_normalized_mql_count
+        mql_count / NULLIF(prior_one_week_mql_count, 0) - 1 AS mql_growth_rate,
+        mql_count * NULLIF(1 - avg_constellation_mql_count_growth, 0) AS normalized_mql_count,
+        mql_growth_rate * NULLIF(1 - avg_constellation_mql_count_growth, 0) AS normalized_mql_growth
+
+        {# MESA MQL Growth #}
     FROM workflow_setup_counts
     LEFT JOIN workflows_created_time_buckets USING (cohort_week)
     LEFT JOIN workflows_enabled_time_buckets USING (cohort_week)
     LEFT JOIN shopify_plan_counts USING (cohort_week)
     LEFT JOIN plan_upgrade_counts USING (cohort_week)
     LEFT JOIN mql_counts USING (cohort_week)
-    LEFT JOIN non_mesa_installs USING (cohort_week)
+    LEFT JOIN decorated_constellation_comparison_groups USING (cohort_week)
 )
 
 SELECT
-    *,
-    LAG(one_week_normalized_mql_count) OVER (ORDER BY cohort_week) AS prior_one_week_normalized_mql_count,
-    1.0 * mql_count / NULLIF(prior_one_week_normalized_mql_count, 0) - 1 AS normalized_growth_rate,
-    1.0 * one_week_normalized_mql_count / NULLIF(prior_one_week_normalized_mql_count, 0) - 1 AS double_normalized_growth_rate
+    *
 FROM final
 WHERE cohort_week < date_trunc('week', CURRENT_DATE())
 ORDER BY cohort_week DESC

models/mesa/staging/stg_step_runs.sql

@@ -24,6 +24,7 @@ decorated_step_runs AS (
         'request',
         'uuid',
         'priority',
+        "__HEVO__SOURCE_MODIFIED_AT",
     ] %}
     SELECT
         _id AS step_run_id,

packages.yml

@@ -15,3 +15,5 @@ packages:
     version: 0.1.1
   - package: dbt-labs/codegen
     version: 0.9.0
+  - package: dbt-labs/dbt_project_evaluator
+    version: 0.6.2