Cohort, RFM and Funnel Analysis: practice questions
400 questions in four levels, all on RetailMart, the practice database of this course. Write every query yourself, get it wrong, read the error, fix it. That is how it sticks.
Answers are coming soon. Post your query in the WhatsApp Community or Discord and we will check it together.
Easy 100 questions Medium 100 questions Hard 100 questions Crazy 100 questions
Core syntax, applied directly
CONCEPTUAL Q1 What is a cohort (group of customers anchored to a common starting date)? Q2 What does signup_month x months-since-signup measure (retention grid)? Q3 What is period_index (months since signup, 0 = signup month)? Q4 What does "Month-1 retention" mean in a cohort? Q5 Why is a cohort triangle the standard shape (cells empty in the future)? Q6 What does RFM stand for (Recency, Frequency, Monetary)? Q7 How does NTILE(5) (Topic 16) produce a 1..5 score per dimension? Q8 Why is "Recency" usually ordered so that low days = high score? Q9 What does a 555 RFM score mean for a customer? Q10 What does "111" RFM score mean? Q11 What is a funnel (counts of users at each step)? Q12 What is "drop-off" between two funnel steps? Q13 What is CLV (customer lifetime value), as a single number? Q14 Difference between MRR-style and per-customer CLV (concept). Q15 Period-over-period growth: MoM / QoQ / YoY - what do they measure? Q16 Why does cohort analysis answer "are new customers staying"? Q17 Why does RFM segmentation drive marketing campaign targeting? Q18 Why does a funnel reveal the worst step in a flow? Q19 What's the anchor date for cohort (typically first order or registration)? Q20 How does a date spine (Topic 23) help with cohort completeness? Q21 Why do we need DISTINCT-customer counts in cohort cells (not order counts)? Q22 What does "active in period N" mean (placed an order in that month)? Q23 Why does a cohort table often show as % retention, not absolute counts? Q24 Why does NTILE require ORDER BY in the OVER() clause? Q25 Name three RetailMart funnels worth measuring (e.g. view->cart->order). COHORT RETENTION Q26 Per customer: signup_month = date_trunc('month', registration_date). Q27 Per order: order_month = date_trunc('month', order_date). Q28 Per (customer, order): months-since-signup = (order_month - signup_month)/30 (concept). Q29 Use AGE or date math (Topic 23) to compute period_index in months. Q30 Cohort sizes: count distinct customers per signup_month. Q31 Active customers in cohort at period_index 0 (signup month). Q32 Active customers per (signup_month, period_index) - cohort grid raw. Q33 Retention %: active / cohort_size per (signup_month, period_index). Q34 Limit to period_index 0..11; tidy long format. Q35 Pivot the long format to wide (signup_month x period_index, Topic 21). Q36 Same as Q35 with retention shown as a percentage. Q37 Month-1 retention per signup_month (a single trend line). Q38 Month-3 retention per signup_month. Q39 Month-6 retention per signup_month. Q40 Compare Month-1 retention 2025 vs 2024 (Topic 23 trend). Q41 Cohort retention per region (join customers->addresses->region). Q42 Cohort retention per acquisition tier (Bronze/Silver/Gold/Platinum). Q43 Cohort sizes vs total customers - sanity reconcile. Q44 Use a date spine (Topic 23) to ensure all (cohort, period) cells exist. Q45 Show the cohort triangle (only past cells filled). Q46 Cohort revenue: revenue per (signup_month, period_index). Q47 Cumulative cohort revenue per period_index (running sum, Topic 17). Q48 Average revenue per customer per cohort per period. Q49 Identify the strongest cohort (highest Month-3 retention). Q50 Identify the weakest cohort (lowest Month-3 retention). RFM SEGMENTATION Q51 Per customer Recency: days since their last order. Q52 Per customer Frequency: count of orders. Q53 Per customer Monetary: SUM(net_total) lifetime. Q54 Recency score: NTILE(5) over -days_since_last_order (low days -> high score). Q55 Frequency score: NTILE(5) over order_count. Q56 Monetary score: NTILE(5) over lifetime_revenue. Q57 Combine into an RFM string score 'RFM' (e.g. '555'). Q58 Total RFM score: R + F + M (1..15). Q59 Top customers by RFM total. Q60 Champions segment: RFM = 555. Q61 At-Risk segment: low R, high F/M (former big spenders going quiet). Q62 New segment: R=5, F=1, M=1 (just started). Q63 Hibernating segment: low R, low F, low M. Q64 Count customers per RFM string score. Q65 Distribution of R score, F score, M score (three histograms). Q66 RFM per region (Topic 21 pivot). Q67 RFM per registration cohort (Topic 23). Q68 Find the median lifetime spend per RFM score (Topic 22). Q69 Average orders per RFM score. Q70 Top-10 customers in Champions by lifetime revenue. Q71 Show how RFM scores correlate with customer tier (Bronze..Platinum). Q72 Per RFM string score: pivot count by region (Topic 21). Q73 Per tier: percent customers in each RFM segment label. Q74 Drift: how many Champions are At-Risk this quarter vs last (Topic 23). Q75 Sanity check: total customers = sum across segments (no double-count). FUNNEL, CLV & PERIOD-OVER-PERIOD Q76 Funnel step 1: distinct customers who viewed any page. Q77 Funnel step 2: distinct customers who placed any order. Q78 Drop-off step 1 -> step 2: % of viewers who ordered. Q79 Funnel: distinct customers who viewed a product page (page_url LIKE '%product%'). Q80 Funnel: distinct customers who added to cart (page_url LIKE '%cart%'). Q81 Funnel: distinct customers who checked out (page_url LIKE '%checkout%'). Q82 3-step funnel counts: product -> cart -> checkout (one wide row). Q83 Drop-off % at each transition (cart/product, checkout/cart). Q84 Sessions instead of customers: per session, did it reach each step? Q85 Funnel per device_type (Topic 21 pivot). Q86 Funnel per month (Topic 23 spine). Q87 CLV per customer: lifetime SUM(net_total) over Delivered orders only. Q88 Average CLV per region (Topic 22 also as median). Q89 CLV per registration cohort (Topic 23). Q90 CLV per RFM segment. Q91 Period-over-period revenue: MoM growth per region (LAG, Topic 18). Q92 PoP: this month vs same month last year (YoY). Q93 PoP: QoQ growth per category (Topic 23 quarter + window). Q94 Active-customer PoP per month per region. Q95 New-customer PoP per month (first-order-month definition). Q96 Cumulative revenue per cohort over months-since-signup. Q97 CLV vs RFM-monetary correlation snapshot. Q98 Cohort retention pivot: signup_month x period_index (Topic 21). Q99 RFM dashboard: per segment count + median CLV (Topic 22). Q100 Exec strip: Month-1 retention, top RFM segment count, funnel drop-off, MoM revenue. Combined ideas, multi-step thinking
CONCEPTUAL Q1 Cohort anchor choice: registration_date vs first_order_date - implications. Q2 Period grain: monthly vs weekly cohorts; tradeoffs. Q3 Why an explicit date spine (Topic 23) prevents missing cells. Q4 Cohort triangle shape: cells above the diagonal are empty (future). Q5 Retention denominator: cohort size at period 0, fixed (not rolling). Q6 Retention vs revenue cohorts - different denominators. Q7 Cumulative cohort LTV vs per-period LTV. Q8 RFM ordering pitfalls: ensure Recency is "smaller-days = higher score". Q9 Ties in NTILE: how Postgres distributes ties (Topic 16). Q10 Named RFM segments: Champions, Loyal, At-Risk, Hibernating, etc. Q11 Segment drift over time: customers moving between segments (Topic 23). Q12 Funnel: customer-level vs session-level - when to choose which. Q13 Drop-off ratio: between consecutive steps; reporting clarity. Q14 Funnel per device/channel/region (Topic 21 pivot). Q15 Topic 22 percentile of CLV per segment - robust segment value. Q16 Topic 22 DISTINCT ON for "latest order per customer" -> R component. Q17 Topic 23 spine to align cohorts cross-month. Q18 Topic 25 MV strategy: which cohort tables benefit. Q19 Topic 26 cleansed inputs: dedup customers before cohort sizing. Q20 Reconciliation: cohort sizes sum to total customers (assertion). Q21 RFM x tier matrix (Topic 21) for marketing teams. Q22 Funnel completeness via spine of steps (Topic 23). Q23 CLV definition choices: gross vs net; included statuses. Q24 PoP zero-division safety (NULLIF). Q25 Reporting Month-1 retention as a single trend (Topic 23). COHORT TRIANGLE & LTV Q26 Per customer signup_month (registration_date -> month). Q27 Per (customer, order) order_month and period_index (months since signup). Q28 Cohort sizes (distinct customers) per signup_month. Q29 Active customers per (signup_month, period_index). Q30 Retention % per (signup_month, period_index). Q31 Limit to period_index 0..11 and pivot wide (Topic 21). Q32 Add a date spine (Topic 23) to guarantee cells. Q33 Cohort revenue per (signup_month, period_index). Q34 Cumulative cohort revenue per period_index (running sum, Topic 17). Q35 Average revenue per active cohort customer per period. Q36 CLV per cohort: cumulative revenue / cohort size at period 0. Q37 Best cohorts by Month-3 retention. Q38 Best cohorts by 12-month cumulative LTV. Q39 Cohort retention by region (Topic 21 pivot). Q40 Cohort retention by acquisition tier. Q41 Cohort retention by signup channel proxy (if any) - concept. Q42 Cohort LTV per region. Q43 Median spend per active cohort customer per period (Topic 22). Q44 P95 spend per active cohort customer per period (Topic 22). Q45 Triangle: signup_month rows x period 0..11 cols, retention values. Q46 Triangle: signup_month rows x period 0..11 cols, LTV values. Q47 Compare 2024 vs 2025 cohorts: Month-1 retention. Q48 Compare 2024 vs 2025 cohorts: 12-month LTV. Q49 Reconcile: cohort sizes sum = total customers. Q50 Save the triangle as a tidy long table (cohort, period, retention, ltv). RFM SEGMENTS & DRIFT Q51 Per customer Recency (days since last order). Q52 Per customer Frequency (order count). Q53 Per customer Monetary (lifetime revenue). Q54 Recency score NTILE(5) over -days_since_last_order. Q55 Frequency score NTILE(5) over order_count. Q56 Monetary score NTILE(5) over lifetime_revenue. Q57 Combine to 'RFM' string score and total R+F+M. Q58 Named segments: Champions, Loyal Customers, Potential Loyalists, New Customers, Promising, Need Attention, About to Sleep, At Risk, Hibernating, Lost. Q59 Apply name-mapping based on score buckets (CASE). Q60 Count customers per named segment. Q61 Median CLV per segment (Topic 22). Q62 P95 CLV per segment. Q63 Per region: counts per segment (Topic 21 pivot). Q64 Per tier: counts per segment. Q65 Drift report: segment this quarter vs last quarter (Topic 23). Q66 Promotion candidates: At Risk + high Monetary. Q67 Reactivation candidates: Hibernating with > 5 historical orders. Q68 New-customer pipeline: count per registration month (Topic 23). Q69 Average days-since-last-order per segment. Q70 Average orders per segment. Q71 Average lifetime revenue per segment. Q72 RFM x region heatmap (Topic 21). Q73 RFM x tier heatmap. Q74 Top-N customers per segment (Topic 16). Q75 Reconcile: customers across segments sum = total customers. FUNNELS & PERIOD-OVER-PERIOD Q76 3-step funnel: product -> cart -> checkout (distinct customers). Q77 4-step funnel: product -> cart -> checkout -> ordered. Q78 Drop-off % per step. Q79 Session funnel via session_id (per session reach). Q80 Funnel per device (Topic 21 pivot). Q81 Funnel per month (Topic 23 spine). Q82 Funnel per region (joins via customers->addresses). Q83 Funnel per registration cohort. Q84 Funnel + median time-between-steps (Topic 22 + Topic 23). Q85 Conversion rate per device per month. Q86 CLV per cohort x per channel proxy. Q87 PoP MoM revenue per region (LAG, Topic 18). Q88 PoP YoY revenue per region. Q89 PoP QoQ revenue per region. Q90 Active-customer PoP per month per region. Q91 New-customer PoP per month. Q92 Reactivated-customer PoP per month. Q93 Funnel conversion drift: WoW shift (Topic 23) per channel. Q94 CLV per RFM segment with MoM trend. Q95 Cohort x CLV table for exec dashboard. Q96 Combined RFM + cohort dashboard per region. Q97 Funnel JSON export (Topic 24) for product team. Q98 Cohort triangle JSON export (Topic 24). Q99 Reconciliation: cohort revenue total = orders revenue (Delivered only). Q100 Exec strip: Month-1 retention by cohort, top RFM segments, funnel drop-off, MoM revenue per region. Interview grade, edge cases
CONCEPTUAL Q1 Cohort anchor: registration vs first_order - implications for retention shape. Q2 Weekly vs monthly cohorts - when noisy/clean. Q3 Cohort triangle storage shape: long vs wide; pivot in BI vs SQL. Q4 Cumulative cohort LTV: when to divide by cohort size vs active size. Q5 RFM ties at percentile boundaries (NTILE vs PERCENTILE_CONT, Topic 22). Q6 Named-segment governance (Champions/Loyal/At-Risk) - rule book. Q7 Drift accounting: customers moving segments quarter-over-quarter. Q8 Funnel: customer reach vs session reach vs per-event count. Q9 Drop-off math: ratio chain vs absolute deltas. Q10 Time-between-steps (median/P95, Topic 22) as a funnel KPI. Q11 Multi-channel/device funnel splits (Topic 21 pivot). Q12 Survivor-bias trap in cohort analysis. Q13 CLV horizon: 12-month vs lifetime; cap and report both. Q14 PoP growth decomposition: new vs returning vs reactivated contribution. Q15 Per-cohort PoP: aging vs same-period-by-cohort. Q16 Spine-driven completeness for cohort and funnel reports (Topic 23). Q17 Topic 26 dependency: cleansed customers/orders/sessions feed the cohort layer. Q18 Topic 25 dependency: which cohort/funnel results should be MVs. Q19 Plan inspection (Topic 19): cohort joins over 150k orders. Q20 Index strategy (Topic 20): customer_id PK + order_date btree, view_timestamp btree. Q21 JSON export (Topic 24) of triangles and funnel payloads. Q22 Segmentation playback: scoring as-of historical points (point-in-time). Q23 Anti-pattern: changing RFM cutoffs without versioning the segment definitions. Q24 Cross-team handoff: marketing wants segments; product wants funnels. Q25 Documenting the cohort/RFM/funnel rules in a runbook. COHORT TRIANGLE + LTV Q26 Monthly cohort grid (signup_month x period 0..11): distinct active customers. Q27 Monthly cohort grid as retention % (active / cohort_size_at_0). Q28 Cumulative cohort LTV per period_index (running sum, Topic 17). Q29 Pivot the LTV grid wide (Topic 21). Q30 Per region cohort grid (join customers->addresses->region). Q31 Per tier cohort grid (Bronze..Platinum). Q32 Weekly cohort grid (signup_week x week 0..12). Q33 Spine-driven completeness (Topic 23) for all (cohort, period) cells. Q34 Cohort size sanity: sum across grid columns equals total. Q35 Best cohort by Month-1 retention. Q36 Best cohort by Month-3 retention. Q37 Best cohort by 12-month LTV. Q38 Worst cohort by Month-3 retention (for diagnostics). Q39 Cohort x region retention heatmap (Topic 21). Q40 Cohort x tier LTV heatmap. Q41 Average revenue per active customer per period (Topic 22 median per cell). Q42 Reactivation rate per cohort (active in period N after gap). Q43 Survival proxy: cohort customers with >=3 active months. Q44 Compare 2024 vs 2025 cohorts: retention shape. Q45 Compare 2024 vs 2025 cohorts: cumulative LTV shape. Q46 JSON cohort export (Topic 24): {cohort, period, retention, ltv} array. Q47 MV candidate (Topic 25): cohort x period retention/LTV - refresh cadence. Q48 Plan check (Topic 19): cohort query on 150k orders + 50k customers. Q49 Index check (Topic 20): orders(order_date, cust_id) helps cohort joins. Q50 Reconciliation: monthly cohort active count vs MAU per month (Topic 23). RFM SEGMENTS + DRIFT Q51 R/F/M raw values per customer (Topic 22 latest order for R). Q52 RFM scores via NTILE(5) on each (Topic 16). Q53 Combine to 'RFM' string + total R+F+M. Q54 Named segments via score buckets (CASE). Q55 Counts per named segment + total customers. Q56 Median & P95 CLV per segment (Topic 22). Q57 Per region segment counts (Topic 21 pivot). Q58 Per tier segment counts. Q59 Per cohort segment counts (cohort_month x segment). Q60 Drift: segment Q1 vs Q2 (transition matrix). Q61 Movers report: Champions in Q1 -> which segments in Q2. Q62 Promotion candidates: At Risk + monetary high. Q63 Reactivation candidates: Hibernating + >=5 historical orders. Q64 New-customer pipeline per month (Topic 23). Q65 Average orders, avg lifetime revenue per segment. Q66 CLV percentile bands per segment (Topic 22). Q67 Top-10 customers per segment by CLV (Topic 16). Q68 Segment x delivery-SLA breach rate (Topic 22). Q69 Segment x support-ticket volume per customer. Q70 JSON RFM segment export (Topic 24). Q71 Point-in-time segmentation: scores as-of last quarter (Topic 22 DISTINCT ON). Q72 Versioned segment rules (v1 vs v2); count divergence. Q73 RFM x cohort retention join: do high-R customers have higher Month-3 retention? Q74 RFM x region anomaly (Topic 22 baseline + segment). Q75 Reconciliation: customers across segments sum = total customers. FUNNELS + GROWTH ATTRIBUTION Q76 4-step funnel: product -> cart -> checkout -> ordered (distinct customers). Q77 Session-level funnel via session_id (per session reach). Q78 Drop-off per step (% lost). Q79 Time-between-steps median/P95 (Topic 22 + Topic 23). Q80 Funnel per device per month (Topic 21 + Topic 23). Q81 Funnel per region per cohort. Q82 Conversion rate per device per month. Q83 PoP funnel: this month vs last month per device. Q84 PoP funnel: same month last year (YoY). Q85 Revenue PoP per region: MoM, YoY, QoQ in one pivot (Topic 21+23). Q86 Growth decomposition per month: new / returning / reactivated revenue. Q87 Active-customer decomposition per month. Q88 Cohort-level PoP: cohort revenue MoM vs aging. Q89 Funnel x RFM segment: do Champions convert better at checkout? Q90 Funnel JSON export (Topic 24) with drop-off and time-between-steps. Q91 CLV per RFM segment with MoM trend. Q92 Triangle pivot to wide (Topic 21) + JSON envelope (Topic 24). Q93 Top movers report: cohorts whose retention shifted > X% (Topic 23). Q94 Plan-check (Topic 19) the heaviest funnel; index strategy (Topic 20). Q95 MV candidates (Topic 25): funnel per month per device; cohort grid. Q96 Reconciliation: funnel-ordered customers = orders.distinct cust_id. Q97 Reconciliation: PoP MoM revenue sum equals monthly_revenue MV (Topic 25). Q98 Cleansed-base (Topic 26) -> cohort layer: deduped customers -> cohort sizes. Q99 Exec dashboard: Month-1 retention per cohort + RFM segment counts + funnel drop-off + MoM revenue per region. Q100 Capstone: cohort triangle (retention+LTV), RFM segments with named labels, 4-step funnel with drop-off and times, MoM/YoY/QoQ per region, and a JSON exec export - one cohesive set of queries. Production scenarios, optimisation
CONCEPTUAL Q1 Architect a growth analytics platform: cleansed -> cohort -> segments -> funnels -> exec. Q2 Cohort anchor choice and impact on retention shape; both grids in parallel. Q3 Weekly cohorts: noise/clean tradeoff; rolling-12-week strategy. Q4 Cumulative LTV per cohort: window functions (17) + spine (23) integration. Q5 Point-in-time RFM scoring vs current-as-of; versioned segment definitions. Q6 Named-segment governance and renamings without breaking dashboards. Q7 Transition matrix (segment Q->Q+1) - interpretation and uses. Q8 Survival models in SQL (concept) for retention curves. Q9 Funnel design: customer vs session vs event grain; choosing per question. Q10 Time-between-steps distribution: P50/P95 (Topic 22) per step. Q11 Attribution by channel/device/region: pivot architectures (Topic 21). Q12 Growth decomposition: new/returning/reactivated revenue + active. Q13 PoP zero/NULL safety; comparing across regimes (seasonality, Topic 23). Q14 Cohort x region x tier 3D slice: pivot architecture limits. Q15 CLV horizons (12-month vs lifetime) and reporting both. Q16 Spine-driven completeness for cohorts/funnels (Topic 23). Q17 MV strategy (Topic 25) for cohort grids and funnel marts. Q18 Plan inspection (Topic 19) of cohort joins; covering indexes (Topic 20). Q19 JSON exec exports (Topic 24): triangles, RFM, funnel envelopes. Q20 Cleansed inputs (Topic 26): why dedup of customers/orders matters. Q21 Idempotent rebuilds of cohort/funnel MVs (refresh DAG). Q22 SLA per insight: freshness, completeness, accuracy. Q23 Cross-team contracts: marketing wants segments; product wants funnels. Q24 Anti-patterns: cohort with a moving cohort_size denominator. Q25 Documentation: cohort/RFM/funnel runbook in MV comments. COHORT + LTV ENGINES Q26 Cohort engine: monthly grid (signup_month x period 0..11) retention. Q27 Cohort engine: cumulative LTV per period (Topic 17 running sum). Q28 Cohort engine: per region/tier/cohort heatmap (Topic 21 pivot). Q29 Cohort engine: weekly grid (signup_week x week 0..12). Q30 Spine-completeness audit: missing cells per cohort (Topic 23). Q31 Best/worst cohort by Month-1/Month-3/12-month LTV (Topic 16 ranks). Q32 Compare 2024 vs 2025 cohorts on both retention and LTV shape. Q33 Reactivation rate per cohort + period (gap + reactivate). Q34 Median revenue per active per cell (Topic 22 in cohort). Q35 Cohort-LTV decomposition: new revenue vs repeat revenue per period. Q36 Cohort x device proxy (concept) using web_events. Q37 Cohort x tier vs region split (Topic 21 multi-dim pivot). Q38 JSON cohort triangle export (Topic 24): {cohort,period,retention,ltv}. Q39 MV mv_cohort_triangle + UNIQUE INDEX(cohort_month, period_index). Q40 MV mv_cohort_ltv + UNIQUE INDEX(cohort_month, period_index). Q41 CONCURRENTLY-refresh both MVs; plan check on the read path (Topic 19). Q42 Cohort engine consumer: exec strip MV joining cohort + RFM. Q43 Cleansed-input dependency (Topic 26): dedup customers before sizing. Q44 Sanity: cohort sizes sum = total customers; LTV sum = total revenue. Q45 Best cohort by 6-month repeat-rate (active in months 1..6). Q46 Cohort-CLV curve per region; identify steepest curves. Q47 Cohort x tier x period - top performers by LTV. Q48 Cohort x channel proxy (web_events) - preliminary attribution. Q49 Cohort export as JSON arrays per region (Topic 24). Q50 Cohort capstone: monthly + weekly grids + LTV + region/tier slice as one mart. RFM PLATFORM & SEGMENT DRIFT Q51 RFM engine: per customer R/F/M raw values (Topic 22 latest order). Q52 NTILE(5) scoring per R/F/M (Topic 16). Q53 Named-segment mapping (Champions, Loyal, At-Risk, ...). Q54 Per region/tier/cohort segment counts (Topic 21). Q55 Median/P95 CLV per segment (Topic 22). Q56 Point-in-time RFM scoring as-of any quarter end. Q57 Transition matrix segment Q->Q+1 (Topic 23). Q58 Movers report: Champions in Q1 -> segments in Q2. Q59 Promotion candidates: At Risk + monetary high; reactivation candidates. Q60 Top movers by CLV per quarter. Q61 RFM x tier x region 3-way matrix (Topic 21). Q62 Versioned segment rules v1/v2 and divergence report. Q63 Segment SLA: % customers stable for >=2 quarters. Q64 JSON RFM export per region (Topic 24). Q65 MV mv_rfm_segments + UNIQUE INDEX(customer_id). Q66 mv_rfm_drift_quarter + UNIQUE INDEX(quarter, customer_id). Q67 mv_segment_ltv_summary + UNIQUE INDEX(segment). Q68 CONCURRENTLY refresh, with plan check (Topic 19). Q69 Segment x delivery SLA breach rate (Topic 22). Q70 Segment x support ticket volume per customer. Q71 Segment x cohort retention correlation. Q72 RFM x month: count per segment per month (Topic 23). Q73 RFM-based marketing list (top movers + promotion candidates). Q74 Plan-check the RFM engine; index strategy (Topic 20). Q75 RFM capstone: scoring + named segments + drift + marketing JSON export. FUNNEL + GROWTH ATTRIBUTION & PRODUCTION MARTS Q76 5-step funnel: view -> product -> cart -> checkout -> ordered (distinct customers). Q77 Session funnel via session_id; reach per session per step. Q78 Drop-off per step + cumulative reach. Q79 Time-between-steps median/P95 per step (Topic 22 + Topic 23). Q80 Funnel per device per month (Topic 21 + Topic 23). Q81 Funnel per region per cohort. Q82 PoP funnel MoM/YoY per device. Q83 Conversion rate per device per month (a moving baseline). Q84 Funnel JSON export (Topic 24) per channel/month with drop-off and times. Q85 MV mv_funnel_steps_monthly + UNIQUE INDEX(month, device). Q86 MV mv_conversion_rates_monthly + UNIQUE INDEX(device, month). Q87 Growth decomposition MV: new/returning/reactivated per month per region. Q88 PoP revenue per region (MoM, YoY, QoQ) in one wide row. Q89 Cohort-aware PoP: cohort revenue MoM vs aging. Q90 Funnel x RFM segment: do Champions convert better? Q91 Reconciliation: funnel ordered-step customers = orders.distinct cust_id. Q92 Reconciliation: PoP MoM revenue total matches mv_daily_revenue (Topic 23/25). Q93 Exec mart: Month-1 retention per cohort + RFM segment counts + funnel drop-off + MoM revenue per region. Q94 JSON exec dashboard export (Topic 24): cohort triangle + RFM segments + funnel + PoP. Q95 Marketing handoff: JSON list of Reactivation Candidates with last order and CLV. Q96 Product handoff: JSON funnel diagnostic with worst-step per channel. Q97 Finance handoff: JSON cohort LTV curves with assumptions in meta. Q98 Drift alerting: cohorts/segments shifting > X% (Topic 23). Q99 End-to-end refresh DAG: cleansed (Topic 26) -> cohort/RFM/funnel MVs (Topic 25) -> JSON exports (Topic 24). Q100 Capstone: production growth analytics platform - cohort triangle (retention+LTV), versioned RFM segments + drift, 5-step funnel with drop-off and times, MoM/YoY/QoQ decomposition, all delivered as MVs with refresh DAG and JSON exec exports.